AI / LLM Engineer — CS graduate from VIT Vellore (2025), building production-grade GenAI systems: agentic AI orchestration, RAG pipelines, LLM fine-tuning (QLoRA/LoRA), and local-first AI platforms. I'm the author of SheshAOS, a governance-first, event-sourced AI operating system written in Rust (9 workspace crates + CLI, 877+ tests, full CI/CD).
🌐 Portfolio · 💼 LinkedIn · 🐙 GitHub · ✉️ Email
| Project | Description | Stack |
|---|---|---|
| SheshAOS | Governance-first, event-sourced AI operating system: policy-enforced agent kernel, append-only audit trail, provider-agnostic LLM streaming (OpenAI/Anthropic), terminal emulation + SSH. 9 workspace crates + CLI, 877+ tests. | Rust, Tokio, LiteLLM, Event Sourcing |
| shesh-ecosystem | Federated, local-first AI body (Brain + Mind + Soma) for CachyOS/Hyprland — orchestrator, manifests, gates, 22 components. | Python, Rust, MCP, Ollama |
| shesh-omniroute | Shesh wrapper for OmniRoute — free open-source AI gateway (291 providers, 90+ free, 500+ models) with automatic tier fallback. | TypeScript, MCP, A2A |
| RAG Service | Production RAG API: hybrid retrieval (dense embeddings + BM25 keyword, fused with Reciprocal Rank Fusion) over ChromaDB. Grounded answers with citations. | Python, FastAPI, ChromaDB |
| LLM Eval Harness | Golden-set evaluation for LLM/RAG systems: faithfulness, answer relevance, correctness with LLM-as-judge + offline fallbacks. CI-ready reports. | Python, LLM-as-Judge, YAML |
| Project | Description | Stack |
|---|---|---|
| Vyākṛti (demo) | Sanskrit-oriented programming language with complete compiler pipeline (lexer → parser → type checker → bytecode) and browser IDE with Monaco, LSP, and REPL. | Rust, React, TypeScript, Axum |
| AIM (demo) | Production-grade attendance management system with Argon2id auth, CSRF protection, Prometheus monitoring, 101 tests, Docker Compose, and CI/CD. | Python, Flask, MySQL, Docker |
| FWRS | Optimizes surplus food allocation to NGOs using 3-stage linear programming (fairness → priority → cost), expiry-aware routing, Folium maps, and a Flask dashboard. | Python, PuLP, Flask, Folium |
| Grievance Portal | Multi-role Laravel 11 app for filing and tracking grievances with role-based access and Docker Compose. | PHP, Laravel 11, MySQL, Docker |
| GameVault | Social gaming marketplace with Next.js 13, Supabase, tag-based search, and a responsive dark-theme UI. | Next.js, TypeScript, Supabase, Tailwind |
| Clinic Ledger OS | Offline-first, voice-assisted clinic ledger and patient memory database for practitioners in India. | Kotlin, Jetpack Compose, MVVM |
| Portfolio (live) | Personal portfolio — Astro + Tailwind CSS 4, dark/light theme, data-driven résumé, and an enterprise-style docs site. | Astro, Tailwind, TypeScript |
AI / LLM: LLM Integration · RAG Pipelines · Agentic AI Frameworks · Prompt & Context Engineering · Multi-Agent Systems · Tool-Using Agents · Fine-tuning (QLoRA/LoRA) · MCP / Tool Protocols · Vector Databases · AI Governance
LLM Ops: Evaluation & Observability · Local LLM Deployment (Ollama, LiteLLM) · OpenAI / Anthropic APIs · Embeddings · Structured Outputs / Function Calling · Streaming Responses
Languages: Python · TypeScript · JavaScript · Rust · PHP · C · C++ · HTML/CSS · Kotlin
Frontend: React · Next.js · Tailwind CSS · Bootstrap · Monaco Editor · React Native
Backend: FastAPI · Flask · Laravel · Node.js · Express · Axum (Rust) · REST APIs · WebSockets · Supabase
Databases: MySQL · MongoDB · Redis · SQLite
DevOps: Docker · GitHub Actions · Nginx · Gunicorn · Prometheus · Linux
Vision / ML: YOLO · OpenCV · Computer Vision · NLP
Industry 5.0 Industrial Automation Trainee — CodenPlay Robotics (Aug–Nov 2025)
- PLC automation (ladder logic, structured text) in CODESYS
- Industrial process simulation in Factory I/O
- Flow-based control/integration in Node-RED
- HMI design, industrial communication, process monitoring
- Email: gagan.jain.se@gmail.com
- LinkedIn: linkedin.com/in/gaganjainse
- Portfolio: gaganjain.vercel.app
Fleet-wide reading compilation: shesh-docs.