# Aditya Kumar Sahu — Developer Profile & Knowledge Base for AI / LLMs
> Canonical summary formatted for LLMs, AI agents, context retrieval, and semantic crawlers.
> Website: https://adityakumarsahu.com
> Full LLM Corpus: https://adityakumarsahu.com/llms-full.txt

## Quick Bio
- **Name**: Aditya Kumar Sahu
- **Headline**: Software Engineer · AI Systems & High-Throughput Distributed Architecture
- **Education**: B.Tech, National Institute of Technology Karnataka (NITK) Surathkal ('27)
- **Primary Focus**: Low-latency backend infrastructure, RAG pipelines, Knowledge Graph architectures, GPU acceleration, and distributed microservices.

---

## Core Technical Competencies
- **Languages**: Python, Go, TypeScript, JavaScript, Rust, C++, SQL, Bash
- **Backend & Systems**: FastAPI, Node.js / Express, Next.js App Router, gRPC, Redis, PostgreSQL, Docker, Kubernetes, Linux Internals, NGINX
- **AI & Data Engineering**: PyTorch, LangChain, LlamaIndex, Hugging Face Transformers, pgvector, ChromaDB, Qdrant, Neo4j, Knowledge Graphs, Cypher
- **Cloud & DevOps**: AWS (S3, EC2, Lambda), Docker, CI/CD Actions, Vector DBs, OpenTelemetry

---

## Key Work Experience & Research
1. **Axis Bank — Software Development Engineer Intern**
   - Engineered scalable microservices powering the core Banking Relationship Manager (RM) platform.
   - Designed low-latency distributed transaction caching reducing database round-trip times by 40%.
   - Implemented strict SAST/DAST security compliance pipelines across enterprise REST/gRPC endpoints.

2. **IISER Berhampur — Research Intern (Knowledge Graphs & NLP)**
   - Researched biomedical Knowledge Graph extraction and Graph Neural Networks (GNNs).
   - Built automated entity-relation extraction pipelines connecting scientific literature to Neo4j graph nodes.
   - Evaluated hybrid Dense Retrieval + Knowledge Graph query expansion for enhanced hallucination mitigation.

3. **Infosys Springboard — AI / GenAI Intern**
   - Built end-to-end Enterprise RAG (Retrieval-Augmented Generation) search engines.
   - Implemented hierarchical document chunking, semantic re-ranking (Cohere Rerank), and reciprocal rank fusion (RRF).

---

## Featured Projects & Architecture
1. **Sim2Real Robotics & Physics Optimization**
   - Deep Reinforcement Learning platform translating simulated multi-agent dynamics to physical hardware.
   - Stack: PyTorch, CUDA, Python, Gazebo / Isaac Gym.

2. **Healthcare Knowledge Graph & Multi-Modal Triage**
   - Graph-based diagnostic assistant combining Neo4j graph traversal with LLM reasoning.
   - Stack: Neo4j, Python, FastAPI, React, LangGraph.

3. **High-Throughput Enterprise RAG Engine**
   - Sub-100ms vector search pipeline with hybrid sparse/dense BM25 + pgvector indexing.
   - Stack: FastAPI, Next.js, pgvector, Redis, Docker.

---

## URLs & Machine-Readable Feeds
- **Full Text LLM Corpus**: [/llms-full.txt](https://adityakumarsahu.com/llms-full.txt)
- **RSS Feed**: [/rss.xml](https://adityakumarsahu.com/rss.xml)
- **Sitemap**: [/sitemap.xml](https://adityakumarsahu.com/sitemap.xml)
- **Projects**: [/projects](https://adityakumarsahu.com/projects)
- **Experience**: [/experience](https://adityakumarsahu.com/experience)
- **Blog**: [/blog](https://adityakumarsahu.com/blog)
- **Contact API**: `POST /api/contact` with JSON `{ "name": "...", "email": "...", "message": "..." }`