# Abhik's Technical Portfolio & Blog

> A comprehensive portfolio and technical blog specializing in Computer Vision, Deep Learning, and AI systems optimization. Features 100+ interactive visualizations, in-depth research paper analysis, and practical ML engineering guides.
> Full version: https://www.abhik.ai/llms-full.txt

## Overview
This website serves as a professional portfolio and technical blog platform built with Next.js 15.5.10 (Pages Router), React, and TailwindCSS. It showcases expertise through in-depth articles, research paper reviews, interactive concept explanations, and practical ML engineering insights.

## Key Information
- **Author**: Abhik Sarkar
- **Role**: Machine Learning Engineer
- **Website**: https://www.abhik.ai
- **GitHub**: @abhiksark
- **Twitter**: @abhiksark
- **LinkedIn**: /in/abhiksark
- **License**: Content may be cited with attribution
- **Version**: 0.38.0
- **Last Updated**: 2026-05-20
- **Primary Focus**: Machine Learning Engineering & Computer Vision
- **Tech Stack**: Next.js 15.5.10 (Pages Router), React 18.2.0, TailwindCSS 3.1.4, MDX
- **Content Types**: Technical Articles (27), Research Papers (31), Interactive Concepts (190), Talks (7)

## Architecture & Technical Details
- **Framework**: Next.js 15.5.10 with **Pages Router** (NOT App Router)
- **Language**: JavaScript/JSX (.js, .jsx files)
- **Styling**: TailwindCSS 3.1.4 exclusively (no CSS-in-JS)
- **Content System**: MDX-based with interactive React components
- **Deployment**: Docker multi-stage builds (output: 'standalone')
- **Analytics**: Vercel Analytics & Speed Insights
- **SEO**: Unified SEO system with dynamic OG images (og.abhik.xyz)
- **Interactive Features**: 100+ custom visualizations with progressive loading
- **Component Patterns**: Compound components, dark mode via classes, responsive canvases

## Directory Structure
- `/src/pages/*` - Next.js pages and routing (Pages Router)
- `/src/components/*` - React components (layouts, articles, concepts, icons)
- `/src/lib/*` - Utility functions (SEO, metadata, content fetching, image optimization)
- `/src/data/*` - Static data (siteMeta, resume, categories)
- `/src/styles/*` - CSS and styling (Tailwind, Prism, typography)
- `/public/*` - Static assets (images, fonts, RSS feeds, sitemaps)

## Important Pages
- [Home](https://www.abhik.ai) - Landing page and introduction
- [About](https://www.abhik.ai/about) - Professional background and expertise
- [Articles](https://www.abhik.ai/articles) - Technical articles and insights
- [Papers](https://www.abhik.ai/papers) - Research paper analysis and reviews
- [Concepts](https://www.abhik.ai/concepts) - ML/AI concept explanations
- [Resume](https://www.abhik.ai/resume) - Professional resume and experience
- [Speaking](https://www.abhik.ai/speaking) - Speaking engagements and talks
- [Uses](https://www.abhik.ai/uses) - Tools and technologies used
- [Consulting](https://www.abhik.ai/consulting) - Consulting services offered
- [Bookmarks](https://www.abhik.ai/bookmarks) - Curated resources and links
- [Talks](https://www.abhik.ai/talks) - Conference talks and presentations

## Research Papers & Analysis (31 Papers)
Comprehensive analysis of foundational and cutting-edge research papers in machine learning and computer vision:

### Foundational Transformer Architecture
- [Attention Is All You Need](https://www.abhik.ai/papers/attention-is-all-you-need) - The Transformer paper (NeurIPS 2017, 180k+ citations): self-attention mechanism and transformer architecture
- [Swin Transformer](https://www.abhik.ai/papers/swin-transformer) - Hierarchical Vision Transformer with shifted window attention (ICCV 2021)
- [Data Movement Transformer](https://www.abhik.ai/papers/data-movement-transformer) - Optimizing transformers through data movement analysis (2021)
- [Optimizing Transformer Inference](https://www.abhik.ai/papers/optimizing-transformer-inference) - Survey of inference optimization: pruning, quantization, distillation (2023)

### Vision Transformers & Image Recognition
- [Vision Transformer (ViT)](https://www.abhik.ai/papers/image-worth-16x16) - "An Image is Worth 16x16 Words": pure transformer for image recognition (2021)
- [ViT Object Detection](https://www.abhik.ai/papers/vit-object-detection) - Plain Vision Transformers as detection backbones (ECCV 2022)

### Object Detection & Computer Vision
- [YOLO](https://www.abhik.ai/papers/yolo) - "You Only Look Once" unified real-time object detection (CVPR 2016)
- [Faster R-CNN](https://www.abhik.ai/papers/faster-rcnn) - Region Proposal Networks for real-time detection (NeurIPS 2015)
- [DETR](https://www.abhik.ai/papers/DETR) - End-to-End Object Detection with Transformers (2020)
- [SURF](https://www.abhik.ai/papers/surf) - Speeded-Up Robust Features for local feature detection (ECCV 2006)

### Segmentation & Advanced Vision
- [SAM](https://www.abhik.ai/papers/sam) - Segment Anything Model with promptable segmentation (2023)

### Multimodal & Vision-Language Models
- [CLIP](https://www.abhik.ai/papers/clip) - Contrastive learning of vision and language on 400M image-text pairs (ICML 2021)
- [BLIP-2](https://www.abhik.ai/papers/blip2) - Efficient vision-language pre-training with frozen encoders and Q-Former (ICML 2023)
- [Visual Instruction Tuning](https://www.abhik.ai/papers/visual-instruction-tuning) - LLaVA: aligning LLMs with visual information (NeurIPS 2023)

### Neural Network Architecture & Optimization
- [Deep Residual Learning](https://www.abhik.ai/papers/deep-residual-learning) - ResNet: skip connections enabling ultra-deep networks (CVPR 2015, 145k+ citations)
- [EfficientNet](https://www.abhik.ai/papers/efficientnet) - Compound scaling for CNNs: depth, width, and resolution (ICML 2019)
- [DeepLearning Go Brr](https://www.abhik.ai/papers/deeplearning-go-brr) - First principles approach to deep learning performance optimization (2022)

### Self-Supervised & Contrastive Learning
- [SimCLR](https://www.abhik.ai/papers/simclr) - Simple framework for contrastive learning of visual representations (ICML 2020)
- [MoCo](https://www.abhik.ai/papers/moco) - Momentum Contrast: dictionary queue for unsupervised visual representation learning (CVPR 2020)
- [BYOL](https://www.abhik.ai/papers/byol) - Bootstrap Your Own Latent: self-supervised learning without negative pairs (NeurIPS 2020)
- [VICReg](https://www.abhik.ai/papers/vicreg) - Variance-Invariance-Covariance regularization for self-supervised learning (ICLR 2022)
- [DINO](https://www.abhik.ai/papers/dino) - Self-distillation with no labels: emergent segmentation in ViT attention maps (ICCV 2021)
- [DINOv2](https://www.abhik.ai/papers/dinov2) - Scaling self-supervised visual features with curated data (LVD-142M) (TMLR 2024)

### Masked Image Modeling & Joint-Embedding
- [MAE](https://www.abhik.ai/papers/mae) - Masked Autoencoders: masking 75% of patches for scalable self-supervised learning (CVPR 2022)
- [BEiT](https://www.abhik.ai/papers/beit) - BERT pre-training for image transformers via discrete visual tokens (ICLR 2022)
- [I-JEPA](https://www.abhik.ai/papers/ijepa) - Joint-Embedding Predictive Architecture for images: abstract feature prediction (CVPR 2023)
- [V-JEPA](https://www.abhik.ai/papers/vjepa) - Video representations by predicting masked regions in latent space (TMLR 2024)
- [V-JEPA 2](https://www.abhik.ai/papers/vjepa2) - Scaling video self-supervised learning to 1M+ hours with mask denoising (arXiv 2025)

### Generative Models & Diffusion
- [DDPM](https://www.abhik.ai/papers/ddpm) - Denoising Diffusion Probabilistic Models: foundation of modern image generation (NeurIPS 2020)
- [Latent Diffusion](https://www.abhik.ai/papers/latent-diffusion) - High-resolution image synthesis in compressed latent space (CVPR 2022, behind Stable Diffusion)
- [Flow Matching](https://www.abhik.ai/papers/flow-matching) - Simplified generative modeling via straight transport paths from noise to data (ICLR 2023)

## Technical Articles (27 Articles)
In-depth technical content focusing on machine learning engineering, system optimization, and practical implementation:

### Video Compression & H.264 Series (4 articles)
- [H.264 Fundamentals](https://www.abhik.ai/articles/h264-fundamentals) - Core pipeline architecture, block-based processing, motion estimation, YUV color space (Part 1/3)
- [H.264 Transform & Quantization](https://www.abhik.ai/articles/h264-transform-quantization) - DCT transforms, quantization, rate-distortion optimization, CABAC vs CAVLC (Part 2/3)
- [H.264 Implementation & Applications](https://www.abhik.ai/articles/h264-implementation-applications) - Profiles/levels, hardware vs software encoding, codec comparison, NVDEC (Part 3/3)
- [H.264 Interactive Guide](https://www.abhik.ai/articles/h264-interactive-guide) - Interactive visualizations of complete H.264 pipeline with 19+ custom components

### C++ Internals Series (3 articles)
- [C++ Compilation Process](https://www.abhik.ai/articles/cpp-compilation-process) - From source to object files: preprocessing, parsing, AST generation, optimization passes
- [C++ Linking In-Depth](https://www.abhik.ai/articles/cpp-linking-in-depth) - Static vs dynamic linking, symbol resolution, GOT/PLT tables, relocations, ELF format
- [C++ Loading and Runtime](https://www.abhik.ai/articles/cpp-loading-runtime) - From executable to process: ELF loading, memory layout, dynamic linker, startup sequence

### Machine Learning Engineering & Optimization (7 articles)
- [Compiling PyTorch Kernel](https://www.abhik.ai/articles/compiling-pytorch-kernel) - torch.compile internals, kernel fusion, TorchInductor, CUDA code generation
- [Kernel Fusion](https://www.abhik.ai/articles/kernel-fusion) - Neural network performance through unified CUDA kernels, memory bandwidth optimization
- [How TensorRT Works](https://www.abhik.ai/articles/how-tensorrt-works) - NVIDIA inference optimization: layer fusion, precision calibration, kernel auto-tuning
- [Quantization Deep Dive](https://www.abhik.ai/articles/quantization-deep-dive) - From FP32 to INT4: QAT vs PTQ, GPTQ, AWQ, SmoothQuant, mixed precision strategies
- [Visualizing YOLOv5](https://www.abhik.ai/articles/visualizing-yolov5) - Visual guide to YOLOv5 architecture, multi-scale detection, anchor boxes
- [GGML Structure](https://www.abhik.ai/articles/ggml-structure) - File structure visualization of GGML format, quantization mapping
- [CPython Internals](https://www.abhik.ai/articles/cpython-internals) - How Python works: bytecode compilation, object model, GIL, memory management

### Computer Vision Deep Dives (3 articles)
- [SAM Multi-Mask Ambiguity](https://www.abhik.ai/articles/sam-multi-mask-ambiguity) - How SAM resolves point prompt ambiguity through three-mask output design and IoU prediction
- [Visualizing YOLOv11](https://www.abhik.ai/articles/visualizing-yolov11) - YOLOv11 loss functions: IoU variants (GIoU, DIoU, CIoU), Distribution Focal Loss, anchor-free detection
- [Numerical Sensitivity](https://www.abhik.ai/articles/numerical-sensitivity) - Why FP16 breaks NAdam: floating-point arithmetic, machine epsilon, catastrophic cancellation

### GPU & System Debugging (3 articles)
- [GPU Boot Errors](https://www.abhik.ai/articles/gpu-boot-errors) - Linux GPU boot errors, initramfs debugging, nouveau vs NVIDIA driver conflicts
- [GPU Xid 31 MMU Faults](https://www.abhik.ai/articles/gpu-xid31-mmu-faults) - Deep dive into NVIDIA Xid 31: GPU virtual memory, page table walk failures, production debugging
- [NVIDIA Xid Error Field Guide](https://www.abhik.ai/articles/nvidia-xid-errors) - Definitive reference for every NVIDIA Xid error code: severity, triage flowcharts, hardware vs software

### PyTorch & Deep Learning Debugging (2 articles)
- [View Size Not Compatible](https://www.abhik.ai/articles/view-size-not-compatible) - PyTorch tensor memory layout, strides, reshape vs view, contiguous operations
- [Registry Pattern](https://www.abhik.ai/articles/registry-pattern) - OpenMMLab's dynamic model loading, decorator-based registration patterns

### Python & Production (1 article)
- [Python Production Logging](https://www.abhik.ai/articles/python-production-logging) - Python's logging module is already a singleton: getLogger, dictConfig, structlog vs stdlib

### Encoding & File Formats (3 articles)
- [Image Encoding](https://www.abhik.ai/articles/image-encoding) - Lossy vs lossless compression: JPEG DCT-based compression vs PNG lossless
- [Text Encoding](https://www.abhik.ai/articles/text-encoding) - ASCII vs UTF-8/16/32 encoding standards, implications for LLM tokenization
- [Magic Numbers](https://www.abhik.ai/articles/magic-numbers) - File signatures and hexadecimal identifiers, focus on JPEG magic number

### Knowledge Management (1 article)
- [Zettel](https://www.abhik.ai/articles/zettel) - Adopting Zettelkasten method for research paper explanations and linking strategy

## Interactive Concepts (190 Concepts)
Interactive explanations with custom visualizations organized into 7 categories:

### Transformers & LLMs (26 concepts)
- [ALiBi](https://www.abhik.ai/concepts/transformers/alibi) - Attention with Linear Biases
- [Alignment Problem](https://www.abhik.ai/concepts/transformers/alignment-problem) - Cross-modal alignment
- [Attention Comparison](https://www.abhik.ai/concepts/transformers/attention-comparison) - MHA vs GQA vs MQA comparison
- [Attention Sinks](https://www.abhik.ai/concepts/transformers/attention-sinks) - Initial token attention phenomena
- [CLS Token](https://www.abhik.ai/concepts/transformers/cls-token) - Classification tokens in transformers
- [Context Windows](https://www.abhik.ai/concepts/transformers/context-windows) - Memory limits of LLMs
- [Cross Attention](https://www.abhik.ai/concepts/transformers/cross-attention) - Attention between different sequences
- [Flash Attention](https://www.abhik.ai/concepts/transformers/flash-attention) - IO-aware exact attention
- [Grouped Query Attention](https://www.abhik.ai/concepts/transformers/grouped-query-attention) - GQA balancing MHA and MQA
- [Hierarchical Attention](https://www.abhik.ai/concepts/transformers/hierarchical-attention) - Multi-scale attention patterns
- [KV Cache](https://www.abhik.ai/concepts/transformers/kv-cache) - Key-value caching for fast inference
- [Linear Attention Approximations](https://www.abhik.ai/concepts/transformers/linear-attention-approximations) - Linear complexity attention
- [Masked Attention](https://www.abhik.ai/concepts/transformers/masked-attention) - Causal masking in autoregressive models
- [Mixture of Experts](https://www.abhik.ai/concepts/transformers/mixture-of-experts) - Sparse MoE models
- [Modality Gap](https://www.abhik.ai/concepts/transformers/modality-gap) - Representation gap between modalities
- [Multi-Query Attention](https://www.abhik.ai/concepts/transformers/multi-query-attention) - MQA for faster inference
- [Multi-Head Attention](https://www.abhik.ai/concepts/transformers/multihead-attention) - Parallel attention mechanisms
- [Positional Embeddings in ViT](https://www.abhik.ai/concepts/transformers/positional-embeddings-vit) - Position encoding for image patches
- [Rotary Position Embeddings](https://www.abhik.ai/concepts/transformers/rotary-position-embeddings) - RoPE for relative positions
- [Scaled Dot-Product Attention](https://www.abhik.ai/concepts/transformers/scaled-dot-product) - Core attention mechanism
- [Scaling Laws](https://www.abhik.ai/concepts/transformers/scaling-laws) - Multimodal scaling
- [Self-Attention in ViT](https://www.abhik.ai/concepts/transformers/self-attention-vit) - Interactive self-attention in Vision Transformers
- [Sliding Window Attention](https://www.abhik.ai/concepts/transformers/sliding-window-attention) - Local attention windows
- [Sparse Attention Patterns](https://www.abhik.ai/concepts/transformers/sparse-attention-patterns) - Efficient sparse attention
- [Tokenization](https://www.abhik.ai/concepts/transformers/tokenization) - Text to token conversion
- [Vision-Language Adapters](https://www.abhik.ai/concepts/transformers/vision-language-adapters) - Adaptation layers

### Deep Learning (34 concepts)
- [Adaptive Tiling](https://www.abhik.ai/concepts/deep-learning/adaptive-tiling) - Dynamic visual token generation
- [Batch Normalization](https://www.abhik.ai/concepts/deep-learning/batch-normalization) - Training acceleration normalization
- [Batch Norm vs Layer Norm: When to Use Which](https://www.abhik.ai/concepts/deep-learning/batch-vs-layer-norm) - BatchNorm normalizes over the batch and spatial axes; LayerNorm normalizes over the channel...
- [Calinski-Harabasz Index](https://www.abhik.ai/concepts/deep-learning/calinski-harabasz) - Variance ratio criterion
- [Collapse Risk](https://www.abhik.ai/concepts/deep-learning/collapse-risk) - Representation collapse in self-supervised learning
- [Contrastive Loss](https://www.abhik.ai/concepts/deep-learning/contrastive-loss) - Similarity-based loss for representations
- [Convolution Operation](https://www.abhik.ai/concepts/deep-learning/convolution-operation) - 2D convolution mechanics
- [Cross-Entropy Loss](https://www.abhik.ai/concepts/deep-learning/cross-entropy-loss) - Classification loss function
- [Davies-Bouldin Index](https://www.abhik.ai/concepts/deep-learning/davies-bouldin) - Worst-case cluster similarity
- [Dilated Convolutions](https://www.abhik.ai/concepts/deep-learning/dilated-convolutions) - Atrous convolutions
- [Dropout](https://www.abhik.ai/concepts/deep-learning/dropout) - Random neuron dropout regularization
- [Emergent Abilities](https://www.abhik.ai/concepts/deep-learning/emergent-abilities) - Capabilities at scale
- [Focal Loss](https://www.abhik.ai/concepts/deep-learning/focal-loss) - Class imbalance handling
- [Gradient Flow](https://www.abhik.ai/concepts/deep-learning/gradient-flow) - Backpropagation dynamics
- [Graph Attention Networks](https://www.abhik.ai/concepts/deep-learning/graph-attention-networks) - GAT with attention
- [Graph Centrality](https://www.abhik.ai/concepts/deep-learning/graph-centrality) - Centrality measures
- [Graph Convolutional Networks](https://www.abhik.ai/concepts/deep-learning/graph-convolutional-networks) - GCN architecture
- [Graph Embeddings](https://www.abhik.ai/concepts/deep-learning/graph-embeddings) - Node and graph embeddings
- [Graph Pooling](https://www.abhik.ai/concepts/deep-learning/graph-pooling) - Graph-level representations
- [He Initialization](https://www.abhik.ai/concepts/deep-learning/he-initialization) - Kaiming initialization for ReLU
- [Internal Covariate Shift](https://www.abhik.ai/concepts/deep-learning/internal-covariate-shift) - Distribution shift problem
- [KL Divergence](https://www.abhik.ai/concepts/deep-learning/kl-divergence) - Distribution distance measure
- [Layer Normalization](https://www.abhik.ai/concepts/deep-learning/layer-normalization) - Layer-wise normalization for transformers
- [MSE vs MAE](https://www.abhik.ai/concepts/deep-learning/mse-mae) - Regression loss functions
- [NAdam Optimizer](https://www.abhik.ai/concepts/deep-learning/nadam) - Nesterov-accelerated Adam
- [Prompt Engineering](https://www.abhik.ai/concepts/deep-learning/prompt-engineering) - Effective prompting strategies
- [Prompt Influence Flow](https://www.abhik.ai/concepts/deep-learning/prompt-influence-flow) - Prompt impact through transformer layers
- [Receptive Field](https://www.abhik.ai/concepts/deep-learning/receptive-field) - Effective receptive field in CNNs
- [Scaling Laws](https://www.abhik.ai/concepts/deep-learning/scaling-laws) - Multimodal scaling
- [Silhouette Score](https://www.abhik.ai/concepts/deep-learning/silhouette-score) - Per-point clustering evaluation
- [Skip Connections](https://www.abhik.ai/concepts/deep-learning/skip-connections) - Residual connections in deep networks
- [VAE Latent Space](https://www.abhik.ai/concepts/deep-learning/vae-latent-space) - Variational autoencoders
- [Visual Complexity Analysis](https://www.abhik.ai/concepts/deep-learning/visual-complexity-analysis) - Image complexity metrics
- [Xavier Initialization](https://www.abhik.ai/concepts/deep-learning/xavier-initialization) - Glorot weight initialization

### Computer Vision (8 concepts)
- [Anchor-Based vs Anchor-Free](https://www.abhik.ai/concepts/computer-vision/anchor-based-vs-anchor-free) - Detection paradigm comparison
- [ASFF](https://www.abhik.ai/concepts/computer-vision/asff) - Adaptive Spatial Feature Fusion
- [Feature Pyramid Networks](https://www.abhik.ai/concepts/computer-vision/feature-pyramid-networks) - Multi-scale feature extraction
- [Modern Object Detection](https://www.abhik.ai/concepts/computer-vision/modern-object-detection) - DETR and transformer-based detectors
- [NAS-FPN](https://www.abhik.ai/concepts/computer-vision/nas-fpn) - Neural Architecture Search for FPNs
- [NMS & Soft-NMS](https://www.abhik.ai/concepts/computer-vision/nms-soft-nms) - Non-Maximum Suppression
- [RoI Pooling](https://www.abhik.ai/concepts/computer-vision/roi-pooling) - RoI Pooling, RoI Align & Deformable RoI
- [Visual Complexity Analysis](https://www.abhik.ai/concepts/computer-vision/visual-complexity-analysis) - Image complexity metrics

### Embeddings & Retrieval (18 concepts)
- [ANN Comparison](https://www.abhik.ai/concepts/embeddings/ann-comparison) - Approximate nearest neighbor methods
- [Binary Embeddings](https://www.abhik.ai/concepts/embeddings/binary-embeddings) - Compact binary vectors
- [BM25 Algorithm](https://www.abhik.ai/concepts/embeddings/bm25-algorithm) - Statistical retrieval
- [Contrastive Learning](https://www.abhik.ai/concepts/embeddings/contrastive-learning) - Contrastive pre-training
- [Cross-Encoder vs Bi-Encoder](https://www.abhik.ai/concepts/embeddings/cross-encoder-vs-bi-encoder) - Architecture comparison
- [Cross-Lingual Alignment](https://www.abhik.ai/concepts/embeddings/cross-lingual-alignment) - Multilingual embeddings
- [Dense Embeddings](https://www.abhik.ai/concepts/embeddings/dense-embeddings) - Vector representations
- [Domain Adaptation](https://www.abhik.ai/concepts/embeddings/domain-adaptation) - Transfer learning for embeddings
- [HNSW Search](https://www.abhik.ai/concepts/embeddings/hnsw-search) - Hierarchical Navigable Small World
- [Hybrid Retrieval Systems](https://www.abhik.ai/concepts/embeddings/hybrid-retrieval-systems) - Dense + sparse retrieval
- [Index Structures](https://www.abhik.ai/concepts/embeddings/index-structures) - Vector index comparison
- [IVF-PQ](https://www.abhik.ai/concepts/embeddings/ivf-pq) - Inverted File with Product Quantization
- [LSH Search](https://www.abhik.ai/concepts/embeddings/lsh-search) - Locality-Sensitive Hashing
- [Matryoshka Embeddings](https://www.abhik.ai/concepts/embeddings/matryoshka-embeddings) - Variable-dimension embeddings
- [Multi-Vector Late Interaction](https://www.abhik.ai/concepts/embeddings/multi-vector-late-interaction) - ColBERT-style retrieval
- [Quantization Effects](https://www.abhik.ai/concepts/embeddings/quantization-effects) - Impact on retrieval quality
- [Sparse vs Dense](https://www.abhik.ai/concepts/embeddings/sparse-vs-dense) - Embedding comparison
- [Vector Quantization](https://www.abhik.ai/concepts/embeddings/vector-quantization) - Compression techniques

### GPU & High-Performance Computing (25 concepts)
- [CUDA Context](https://www.abhik.ai/concepts/gpu-computing/cuda-context) - GPU resource management
- [CUDA Context vs CUDA Stream](https://www.abhik.ai/concepts/gpu-computing/cuda-context-vs-streams) - A CUDA context is a per-device container of GPU state; a CUDA stream is an in-order...
- [CUDA MPS](https://www.abhik.ai/concepts/gpu-computing/cuda-mps) - Multi-Process Service for GPU sharing
- [CUDA Streams: Asynchronous Execution and Concurrency](https://www.abhik.ai/concepts/gpu-computing/cuda-streams) - A CUDA stream is a queue of GPU operations that execute in order
- [Distributed Parallelism](https://www.abhik.ai/concepts/gpu-computing/distributed-parallelism) - Tensor, pipeline, and data parallelism
- [Flynn's Classification](https://www.abhik.ai/concepts/gpu-computing/flynns-classification) - Taxonomy of computer architectures
- [HBM Memory](https://www.abhik.ai/concepts/gpu-computing/hbm-memory) - High Bandwidth Memory
- [HPC Performance Optimization](https://www.abhik.ai/concepts/gpu-computing/hpc-performance-optimization) - Scaling, profiling, and tuning
- [Kubernetes Operator](https://www.abhik.ai/concepts/gpu-computing/kubernetes-operator) - NVIDIA GPU orchestration
- [Memory Hierarchy](https://www.abhik.ai/concepts/gpu-computing/memory-hierarchy) - GPU memory levels and optimization
- [MPI Fundamentals](https://www.abhik.ai/concepts/gpu-computing/mpi-fundamentals) - Message passing for distributed computing
- [Multi-GPU Communication](https://www.abhik.ai/concepts/gpu-computing/multi-gpu-communication) - NVLink, PCIe, and NCCL
- [NCCL Communication](https://www.abhik.ai/concepts/gpu-computing/nccl-communication) - Multi-GPU collective communication
- [NVIDIA Device Files](https://www.abhik.ai/concepts/gpu-computing/nvidia-device-files) - Device file system
- [NVIDIA Persistence Daemon](https://www.abhik.ai/concepts/gpu-computing/nvidia-persistence-daemon) - GPU daemon management
- [OpenMP](https://www.abhik.ai/concepts/gpu-computing/openmp) - Shared-memory parallel programming
- [Page Migration](https://www.abhik.ai/concepts/gpu-computing/page-migration) - CPU-GPU page migration
- [Shared Multiprocessor](https://www.abhik.ai/concepts/gpu-computing/shared-multiprocessor) - SM architecture
- [Slurm Accounting](https://www.abhik.ai/concepts/gpu-computing/slurm-accounting) - Accounting and resource tracking
- [Slurm Backfill](https://www.abhik.ai/concepts/gpu-computing/slurm-backfill) - Backfill scheduling for small jobs
- [Slurm Fundamentals](https://www.abhik.ai/concepts/gpu-computing/slurm-fundamentals) - Job scheduling on HPC clusters
- [Slurm GPU Allocation](https://www.abhik.ai/concepts/gpu-computing/slurm-gpu-allocation) - GPU allocation for distributed training
- [Slurm Resource Management](https://www.abhik.ai/concepts/gpu-computing/slurm-resource-management) - Resource management and job priority
- [Tensor Cores](https://www.abhik.ai/concepts/gpu-computing/tensor-cores) - Mixed-precision acceleration
- [Unified Memory](https://www.abhik.ai/concepts/gpu-computing/unified-memory) - Managed memory

### Systems & Architecture (48 concepts)
- [Boot Process](https://www.abhik.ai/concepts/systems/boot-process) - Linux boot sequence from power-on
- [Btrfs Filesystem](https://www.abhik.ai/concepts/systems/btrfs-filesystem) - B-tree copy-on-write filesystem
- [cgroups](https://www.abhik.ai/concepts/systems/cgroups) - Resource limits for processes
- [Client-Server Communication](https://www.abhik.ai/concepts/systems/client-server-communication) - Polling vs WebSockets
- [Containers](https://www.abhik.ai/concepts/systems/containers) - From primitives to Docker
- [Copy-on-Write](https://www.abhik.ai/concepts/systems/copy-on-write) - CoW mechanism
- [CPU Cache Lines](https://www.abhik.ai/concepts/systems/cpu-cache-lines) - Cache line structure and simulator
- [CPU Optimization](https://www.abhik.ai/concepts/systems/cpu-optimization) - CPU performance and optimization techniques
- [CPU Pipeline Detailed](https://www.abhik.ai/concepts/systems/cpu-pipeline-detailed) - Pipeline stage architecture
- [CPU Pipelines](https://www.abhik.ai/concepts/systems/cpu-pipelines) - Instruction pipelining and branch prediction
- [Ext4 Filesystem](https://www.abhik.ai/concepts/systems/ext4-filesystem) - The Linux workhorse filesystem
- [FAT Filesystems](https://www.abhik.ai/concepts/systems/fat-filesystems) - FAT32 & exFAT universal filesystems
- [Filesystem Integrity](https://www.abhik.ai/concepts/systems/filesystem-integrity) - Silent corruption detection
- [Filesystem Journaling](https://www.abhik.ai/concepts/systems/filesystem-journaling) - Write-ahead logging
- [Filesystem Snapshots](https://www.abhik.ai/concepts/systems/filesystem-snapshots) - Point-in-time copies
- [Filesystems Overview](https://www.abhik.ai/concepts/systems/filesystems-overview) - Filesystem types comparison
- [FUSE Filesystem](https://www.abhik.ai/concepts/systems/fuse-filesystem) - Filesystem in userspace
- [GPU Containers](https://www.abhik.ai/concepts/systems/gpu-containers) - How Docker works with GPUs
- [Hazard Detection](https://www.abhik.ai/concepts/systems/hazard-detection) - Pipeline dependencies and solutions
- [How RAM Works](https://www.abhik.ai/concepts/systems/how-ram-works) - DRAM fundamentals
- [Init Systems](https://www.abhik.ai/concepts/systems/init-systems) - SysV to systemd
- [Initramfs Boot Process](https://www.abhik.ai/concepts/systems/initramfs-boot-process) - Initial RAM filesystem
- [Inodes](https://www.abhik.ai/concepts/systems/inodes) - File metadata structure
- [Kernel Architecture](https://www.abhik.ai/concepts/systems/kernel-architecture) - Linux kernel design
- [Kernel Modules](https://www.abhik.ai/concepts/systems/kernel-modules) - Loadable modules
- [Long Polling](https://www.abhik.ai/concepts/systems/long-polling) - Hanging request pattern
- [Memory Access Patterns](https://www.abhik.ai/concepts/systems/memory-access-patterns) - Sequential vs strided access
- [Memory Controllers](https://www.abhik.ai/concepts/systems/memory-controllers) - RAM management
- [Memory Interleaving](https://www.abhik.ai/concepts/systems/memory-interleaving) - Parallel memory access
- [Memory Management](https://www.abhik.ai/concepts/systems/memory-management) - Python memory allocator
- [Mount Options](https://www.abhik.ai/concepts/systems/mount-options) - Filesystem mount flags
- [Namespaces](https://www.abhik.ai/concepts/systems/namespaces) - Foundation of container isolation
- [Networking Stack](https://www.abhik.ai/concepts/systems/networking-stack) - TCP/IP stack
- [NTFS Filesystem](https://www.abhik.ai/concepts/systems/ntfs-filesystem) - Master File Table
- [NUMA Architecture](https://www.abhik.ai/concepts/systems/numa-architecture) - Non-Uniform Memory Access
- [NVIDIA Modeset](https://www.abhik.ai/concepts/systems/nvidia-modeset) - Kernel modesetting
- [Process Management](https://www.abhik.ai/concepts/systems/process-management) - Fork, exec, and beyond
- [RAID Storage](https://www.abhik.ai/concepts/systems/raid-storage) - Redundant arrays
- [Short Polling](https://www.abhik.ai/concepts/systems/short-polling) - Periodic request pattern
- [SoA vs AoS](https://www.abhik.ai/concepts/systems/soa-vs-aos) - Data layout optimization
- [System Calls](https://www.abhik.ai/concepts/systems/system-calls) - User-kernel interface
- [TCP/IP](https://www.abhik.ai/concepts/systems/tcp-ip) - Internet protocol suite
- [Transparent Huge Pages](https://www.abhik.ai/concepts/systems/transparent-huge-pages) - Large page TLB optimization
- [Virtual Memory](https://www.abhik.ai/concepts/systems/virtual-memory) - Virtual address space and TLB
- [Wayland vs X11](https://www.abhik.ai/concepts/systems/wayland-x11) - Display server architecture
- [WebSocket](https://www.abhik.ai/concepts/systems/websocket) - Real-time bidirectional communication
- [XFS Filesystem](https://www.abhik.ai/concepts/systems/xfs-filesystem) - High-performance parallel FS
- [ZFS Filesystem](https://www.abhik.ai/concepts/systems/zfs-filesystem) - Zettabyte filesystem

### Language & Framework Internals (31 concepts)
- [AST Parsing](https://www.abhik.ai/concepts/language-internals/ast-parsing) - Abstract syntax tree generation
- [Asyncio Event Loop](https://www.abhik.ai/concepts/language-internals/asyncio-event-loop) - Asynchronous I/O
- [Bytecode Compilation](https://www.abhik.ai/concepts/language-internals/bytecode-compilation) - CPython compiler pipeline
- [Compilation](https://www.abhik.ai/concepts/language-internals/compilation) - C++ compilation overview
- [DataParallel vs DDP](https://www.abhik.ai/concepts/language-internals/data-parallel) - Parallel training strategies
- [DataLoader Pipeline](https://www.abhik.ai/concepts/language-internals/dataloader-pipeline) - Data loading architecture
- [Dynamic Linking](https://www.abhik.ai/concepts/language-internals/dynamic-linking) - Runtime linking
- [Garbage Collection](https://www.abhik.ai/concepts/language-internals/garbage-collection) - Reference counting + cyclic GC
- [Global Interpreter Lock](https://www.abhik.ai/concepts/language-internals/global-interpreter-lock) - GIL mechanics
- [Green Threads vs OS Threads](https://www.abhik.ai/concepts/language-internals/green-threads-vs-os-threads) - Threading models
- [Linking](https://www.abhik.ai/concepts/language-internals/linking) - Static and dynamic linking
- [Loading](https://www.abhik.ai/concepts/language-internals/loading) - From ELF to running process
- [Memory Management](https://www.abhik.ai/concepts/language-internals/memory-management) - Python memory allocator
- [Memory & RAII](https://www.abhik.ai/concepts/language-internals/memory-raii) - Resource management
- [Modern C++ Features](https://www.abhik.ai/concepts/language-internals/modern-cpp-features) - C++11/14/17/20 features
- [num_workers](https://www.abhik.ai/concepts/language-internals/num-workers) - Worker process configuration
- [Object Model](https://www.abhik.ai/concepts/language-internals/object-model) - PyObject structure
- [OOP Inheritance](https://www.abhik.ai/concepts/language-internals/oop-inheritance) - Object-oriented inheritance
- [Optimization](https://www.abhik.ai/concepts/language-internals/optimization) - Compiler optimizations
- [Pin Memory](https://www.abhik.ai/concepts/language-internals/pin-memory) - Pinned memory and DMA transfers
- [Pointers and References](https://www.abhik.ai/concepts/language-internals/pointers-references) - Memory addressing fundamentals
- [Preprocessor](https://www.abhik.ai/concepts/language-internals/preprocessor) - Preprocessor directives
- [Python Optimization](https://www.abhik.ai/concepts/language-internals/python-optimization) - Performance techniques
- [Shared Memory](https://www.abhik.ai/concepts/language-internals/shared-memory) - Python shared memory for multiprocessing
- [Slots Optimization](https://www.abhik.ai/concepts/language-internals/slots-optimization) - __slots__ memory optimization
- [Smart Pointers](https://www.abhik.ai/concepts/language-internals/smart-pointers) - RAII and modern memory management
- [Stack vs Heap](https://www.abhik.ai/concepts/language-internals/stack-heap) - Memory allocation
- [Symbol Resolution](https://www.abhik.ai/concepts/language-internals/symbol-resolution) - How the linker connects code
- [Templates & STL](https://www.abhik.ai/concepts/language-internals/templates-stl) - Generic programming
- [Thread Safety](https://www.abhik.ai/concepts/language-internals/thread-safety) - Concurrent programming fundamentals
- [Virtual Tables & Inheritance](https://www.abhik.ai/concepts/language-internals/virtual-tables-inheritance) - Polymorphism implementation

## Talks & Presentations (7 Talks)
- [ArrPy: Array You Fast Enough?](https://www.abhik.ai/talks/arrpy-array-you-fast-enough) - 3-hour workshop rebuilding NumPy from scratch (PyCon India 2025)
- [GPU Programming 101 in Python](https://www.abhik.ai/talks/gpu-programming-101-in-python) - Hands-on Triton workshop: GPU kernel programming in Python (BangPypers December 2025)
- [Understanding Multimodal Models](https://www.abhik.ai/talks/multimodal-models) - CLIP and Gemini architecture, robotics applications (Build with AI - Bangpypers, March 2025)
- [Speeding up Python with Cython](https://www.abhik.ai/talks/speeding-up-python-with-cython) - Performance optimization using Cython (PyCon India/Japan 2024)
- [Rolling with Python: Intro to Python Wheels](https://www.abhik.ai/talks/rolling-with-python-intro-to-python-wheels) - Python packaging and wheel internals (BangPypers 2024)
- [The Age of Digital Da Vinci](https://www.abhik.ai/talks/the-age-of-digital-da-vinci) - Image generation: VAEs, GANs, Stable Diffusion (Devfest 2023)
- [Introduction to Machine Learning](https://www.abhik.ai/talks/introduction-to-machine-learning) - ML fundamentals with real-world examples (KJ Somaiya Techfest 2019)

## Interactive Visualizations (100+ Components)
Major interactive suites include:
- **H.264 Video Compression**: 19+ components covering motion estimation, DCT transforms, quantization, entropy coding
- **TensorRT Optimization**: Layer fusion demos, precision calibration visualizers
- **Quantization Demos**: FP32/INT8/INT4 interactive comparisons
- **Memory Systems**: Cache line simulations, virtual memory visualizations
- **GPU Architecture**: CUDA context management, memory hierarchy
- **Attention Mechanisms**: Self-attention, multi-head, cross-attention visualizations
- **Graph Neural Networks**: GCN and GAT interactive demos
- **CPU Pipelines**: Pipeline stage visualization with hazard detection
- **Self-Supervised Learning**: SimCLR, MoCo, BYOL, DINO, MAE pipeline demos
- **Diffusion Models**: DDPM, Latent Diffusion, Flow Matching demos

## API Endpoints & Services

### RSS Feeds (Multi-format)
- Atom: https://www.abhik.ai/rss/atom.xml
- JSON Feed: https://www.abhik.ai/rss/feed.json
- RSS 2.0: https://www.abhik.ai/rss/feed.xml

### SEO & Discovery
- Sitemap: https://www.abhik.ai/sitemap.xml
- Dynamic OG images: https://og.abhik.xyz/api/og
- Tag pages: /articles/tags/[tag], /papers/tags/[tag], /concepts/tags/[tag]

## Metadata
```json
{
  "name": "Abhik Sarkar",
  "role": "Machine Learning Engineer",
  "version": "0.38.0",
  "lastUpdated": "2026-05-20",
  "contentTypes": {
    "articles": 27,
    "papers": 31,
    "concepts": 190,
    "talks": 7,
    "interactiveComponents": 100
  },
  "categories": 7,
  "specialties": [
    "Computer Vision",
    "Deep Learning",
    "ML Engineering",
    "Video Compression (H.264)",
    "GPU Optimization",
    "Performance Tuning",
    "Systems Programming (C++, Python)",
    "Inference Optimization",
    "Quantization",
    "Multimodal AI",
    "Self-Supervised Learning",
    "High Performance Computing"
  ]
}
```

## Additional Resources
- [GitHub Profile](https://github.com/abhiksark) - Open source contributions and projects
- [LinkedIn Profile](https://linkedin.com/in/abhiksark) - Professional experience
- [Twitter/X](https://twitter.com/abhiksark) - Technical updates and insights