AI/ML Engineer | NLP | Generative AI | LLMs
Building practical AI systems that solve real problems.
| Project | Description | Tech Stack | Demo |
|---|---|---|---|
| Intellecta | Autonomous multi-agent research engine with 4 specialized agents | LangGraph, LangChain, Ollama, Streamlit | Live Demo |
| LexAI | Legal document intelligence: RAG Q&A, risk scoring, compliance | Groq, FAISS, LangChain, Streamlit | Live Demo |
| Job Extractor | Fine-tuned Llama-3.2-3B for structured job parsing (QLoRA, GGUF) | Unsloth, Ollama, FastAPI, Streamlit | Live Demo |
| Movie Recommender | Content-based recommender using TF-IDF + Cosine Similarity | Scikit-learn, Streamlit, Pandas | Live Demo |
Languages:
Python SQL JavaScript TypeScript
AI/ML:
PyTorch TensorFlow Scikit-learn LangChain LangGraph Hugging Face Ollama Unsloth QLoRA RAG FAISS
LLM & GenAI:
Llama 3.x Mistral Groq OpenAI API Anthropic API Prompt Engineering Fine-tuning Quantization
MLOps & Deployment:
FastAPI Streamlit Docker GitHub Actions Hugging Face Spaces Hugging Face Hub MLflow
Data & Tools:
Pandas NumPy Jupyter Git Linux VS Code
I actively maintain a structured knowledge base and daily engineering log tracking practical implementations and system architectures:
- 🤖 Multi-Agent Architectures — LangGraph state graphs, supervisory routing, reflection loops.
- 🔬 LLM Fine-Tuning Playbook — Math and code for LoRA, QLoRA, and NF4 quantization.
- 📚 Production RAG Patterns — Hybrid search (BM25 + Dense) and Reciprocal Rank Fusion.
- ⚡ MLOps Model Registry — Experiment tracking and signature governance with MLflow.
- 🐍 High-Performance Python — Zero-copy
memoryview,__slots__, and async pipelines. - 📅 Daily Learning Journal — Daily problem-solving logs and engineering notes.
- 🔬 Multi-Agent Systems — LangGraph orchestration, agent communication patterns
- 🧠 LLM Fine-tuning — QLoRA, Unsloth, efficient quantization (GGUF, AWQ)
- 📚 RAG & Knowledge Retrieval — FAISS, hybrid search, reranking, evaluation
- ⚡ MLOps — CI/CD for ML, model registry, monitoring, automated retraining
