About
ML Engineer building
production AI systems
I'd rather ship code than write about it. I build ML tools, break things in production, and occasionally publish papers to prove I'm not just hacking.
// try the konami code
After finishing my B.Tech in Computer Science (AI & ML) at VIT Chennai, I'm now pursuing a Master of AI at RMIT in Melbourne. Along the way I've shipped production ML models at Sterling Software, built LLM-powered commerce integrations at Red Augment, and published medical imaging research in Springer Nature.
My focus is on practical AI: building tools like doc-scraper (108+ stars) for LLM pipelines, creating AI-powered code review systems, and developing computer vision models for medical diagnostics.
I'm open to collaborative opportunities in AI research and engineering. If you're working on something interesting, let's talk.
Currently exploring: agentic AI
Experience
Backend API Developer
Red Augment, Singapore (Remote) • Feb – Jul 2026
- Built the subscription commerce stack for WineSpeak.ai (now RedConcierge), an LLM-powered winery chatbot, integrating Shopify, Awtomic, and WineHub across winery deployments with catalogue sync, cart-via-chat, checkout, and subscription management.
- Fixed slow time-to-first-token and bursty streaming after the backend's async migration by running classification, retrieval, and prompt construction concurrently with unbuffered token streaming, and cut member-portal subscription lookups about 2.3x (5.7 s to 2.5 s) by parallelising Shopify and WineHub calls.
ML Engineer Intern
Sterling Software, Chennai • May – Jun 2025
- Built and deployed the company's first ML-based API monitoring system in 6 weeks, processing millions of log rows daily to flag anomalies before they reached production.
- Developed an API latency predictor (R² > 0.85) on Dask and Polars log pipelines.
Open Source Contributor
Python and Go projects • Sep 2026 – Present
- Merged fixes into docling (HTML list and table parsing), Unstructured (table captions), Microsoft's PyRIT (AI red-teaming video converter), and trafilatura (link-text escaping).
- Narrowed OIDC reconcile watches in Stacklok's ToolHive Kubernetes operator for MCP servers, and added Wikisource as a new source to openZIM's Project Gutenberg scraper (feature with tests).
Education
Master of Artificial Intelligence
RMIT University, 2026 – 2028
B.Tech CS (AI & ML)
VIT Chennai, 2026
Research
2 research papers
Medical imaging published in Signal, Image and Video Processing (Springer Nature); computer networks under revision at Peer-to-Peer Networking and Applications (Springer)
Technologies
Languages
- Python
- Go
- SQL
ML & AI
- PyTorch
- TensorFlow
- scikit-learn
- CV
- LLMs
Data
- Dask
- Polars
- Pandas
- PostgreSQL
- MongoDB
Tools
- Docker
- FastAPI
- Django
- Git
- Actions