I'm Haadhi Irfan, an AI Engineer specialising in multi-agent LLM systems, RAG pipelines, and production AI microservices.

• 🎓 MSc Artificial Intelligence & Machine Learning — University of Birmingham (Distinction)
• 🎓 B.Tech (Hons.) Ocean Engineering & Naval Architecture, Spec. AI — IIT Kharagpur (Distinction)
• 🧠 Passionate about agentic systems, LLM integration, and scalable AI infrastructure
• 🧩 Strong foundations in Mathematics, Data Structures, and Algorithms
• ⚙️ Currently building production multi-agent microservices with LangGraph, LangChain, FastAPI, and Groq
• 💬 Ask me about: Python, Multi-Agent Systems, RAG, Prompt Engineering, or LLM fine-tuning
• ♟️ Fun fact: I love playing chess and explaining mathematical concepts to students
• 📫 Reach me: haadhi.irfan@gmail.com

StAnify is a modular multi-agent AI tutoring system built as my Master's Thesis at the University of Birmingham.

• Dedicated agents for analogy generation, narrative building, evaluation, and image critique
• RAG pipeline over structured educational content using Qdrant vector search + vocabulary scaffolding
• Integration of Stable Diffusion XL and OpenAI GPT-4o/5 via CrewAI multi-agent workflows
• Reduced teacher preparation time from 50–80 mins → ~3 mins (17–25× faster) at ~$0.15 per concept 🏆 Preferred in 65% of blinded A/B tests vs GPT-4 — outperforming on clarity (+1.59) and engagement (+1.51)
  🧩 Tech stack: Python · CrewAI · OpenAI API · Stable Diffusion XL · Qdrant · Streamlit · FastAPI

• 🤖 AI Engineer — Aspect Maintenance, London (March 2026 – Present)
  Built a production multi-agent LLM microservice (LangGraph · LangChain · FastAPI · Groq Llama-3.3-70B) for natural language querying of engineer KPIs across 18 metrics; integrated Salesforce CRM and Webfleet as live data sources; enforced multi-layer RBAC across 3 access tiers
• 🧑‍💻 Full-Stack Software Engineer — Einstein11Plus, Birmingham (Jan 2026 – March 2026)
  Engineered a webhook-driven booking and automation engine integrating Shopify with internal systems; reduced manual operations by ~80% and response times by ~30× (6s → <200ms)
• 💡 AI Consultant — Muslim Hands (NGO), Nottingham (March 2025 – June 2025)
  Delivered an AI Use Case & Risk Report and executive AI Training Plan; designed GDPR-compliant AI data strategy and prompt engineering workshops for non-technical stakeholders
• 🌊 AI Engineer (Micro-Internship) — University of Birmingham × Neptune Submerge (July 2025 – August 2025)
  Built MarineBot — RAG + LLM backend (FastAPI · LLaMA 3 · MiniLM · ChromaDB) with Whisper STT and voice interaction; achieved <5s end-to-end latency
• 🛰️ Network Operations Intern — Sterlite Technologies, Gurugram (May 2023 – July 2023)
  Integrated an LLM assistant into enterprise NMS via continued pretraining (~200M samples), SFT (~10k pairs), and RLHF reward modelling (~30k preference pairs); ~35% throughput improvement on data pipelines

• 🥇 Rank 10 — IndoML 2023 (IIT Bombay): multilingual intent detection, fine-tuned BERT, 0.86 leaderboard score
• 🥈 Rank 11 — Kharagpur Data Science Hackathon 2024 (IIT KGP): BTC-USD trading pipeline with time-series, ML, DL & RL
• 🎯 Finalist — Master's Consultancy Challenge 2025 (University of Birmingham × Turner & Townsend)
• 📊 Top 1% of 1.4M+ candidates — JEE Main & Advanced national engineering entrance examination

AI & LLM

ML & Deep Learning

Backend & APIs

Vector & Data

Cloud & DevOps

Frontend

✨ "Curiosity drives my code. Every project is an opportunity to learn, break, and rebuild something better."