| Project | What it does | Stack |
|---|---|---|
| Full-ML-Pipeline | End-to-end ML pipeline: ingestion → training → FastAPI serving | Python · Scikit-learn · Docker |
| Mini_LLM | 3-layer APP with Transformer built from scratch to understand LLMs at the weight level | FastAPI, React, Python · PyTorch |
| StudyPilot | Full-stack Agentic learning platform with RAG & ML | Agents, PostgreSQL/pgvector, Docker, FastAPI |
- 🎓 2nd year, AI Engineering — TUS Athlone (BSc Software Design with AI for Cloud Computing)
- 🏭 AI Intern @ Ericsson — cloud/platform engineering, Cloud RAN test migration, automation tooling
- 🥈 2nd Place, Claude Hackathon @ TCD — built Voxify, a student feedback platform
- 📊 Kaggle Dataset Expert — ranked #194 of 8,000+ · top 2.4% globally
- 🎙️ Class Representative — elected voice for the AI & CLoud Engineering cohort at TUS
- 🧬 Built a transformer from scratch — Mini_LLM
focus = ["LLMs", "ML Engineering", "Data Pipelines", "MLOps"]
languages = ["Python", "JavaScript / TypeScript", "Java"]
ml_stack = ["PyTorch", "Scikit-learn", "NumPy", "Pandas", "Hugging Face"]
llm_tools = ["Anthropic SDK", "OpenAI SDK", "Ollama", "RAG", "Multi-agent systems"]
infra = ["Docker", "Git", "FastAPI", "PostgreSQL"]
learning = ["QLoRA fine-tuning", "Vector DBs", "MLOps pipelines"]I build things to understand how they break. Currently focused on: making ML systems production-ready
Open to research collaborations, internship opportunities, and honest technical conversations.