Headline

AI Systems Engineer | LLMs • NLP • MLOps | PyTorch • HuggingFace • FastAPI • React

About

I mostly spend my time trying to persuade computers to be slightly more intelligent than they were the day before. I enjoy building projects where machine learning meets real software. This ranges from LLM tools and ML pipelines to AI-powered web applications. My goal is simple: build interesting ideas that actually run in production (or at least run without crashing… most of the time). I’m particularly interested in how AI systems interact with real products and developer infrastructure. Things I usually work with: Python • PyTorch • HuggingFace • FastAPI • React • Node.js Things I enjoy doing: • Building LLM-powered applications • Designing ML pipelines that won’t break at 2 AM • Turning research ideas into working products • Occasionally arguing with GPUs You can find my experiments here: 🐙 GitHub: github.com/larrikin-coder 🤗 Hugging Face: huggingface.co/larrikin-coder If you're building interesting AI systems or slightly overambitious side projects, we’ll probably get along.

Location

Ghaziabad

AI stack

claude code

claude

cursor

langchain

langgraph

langsmith

Work experience

Python Intern at Techolution, May 2026 - Present · 5 mos, Hyderabad, Telangana, India

Software Developer Intern at Oil and Natural Gas Corporation Limited (ONGC), Jul 2024 - Aug 2024 · 2 mos, Delhi, India, Built a scalable NLP pipeline in Python to preprocess and clean the FlanV2-100k dataset for improved LLM training. Implemented TF-IDF and cosine similarity based deduplication to remove redundancy across 100k+ samples. Fine-tuned LLaMA2-7B using LoRA with 4-bit quantization and published curated datasets and models on Hugging Face.

Education

Vellore Institute of Technology

Public links

github: https://github.com/larrikin-coder