Mangadoddi Srikar Vardhan
AI/ML Research Engineer at Petavue. I care about two things: understanding how models work on the inside, and turning that understanding into software people can actually use.
I am a machine learning engineer and researcher based in Chennai. I finished my B.Tech in Computer Science at NIT Silchar in 2026, and over my degree I published six peer-reviewed papers, two of them as first author, on LLM interpretability, model compression, and AI-generated text detection.
My research side leans toward interpretability and efficiency: figuring out which parts of a model matter and using that to compress it without losing accuracy. My builder side is broader. I have shipped a desktop companion to the Microsoft Store, a speech-to-speech pipeline with voice cloning and lip-sync, RAG assistants, an agentic tool for spreadsheets, and a Transformer translation model written from scratch.
Right now I work full time at Petavue on LLM interpretability, compression, and NLP. I also take on a small amount of freelance work when a problem is interesting.
Research & ML
Engineering
I build
Get in touch
Open to ML engineering roles, research collaborations, and freelance builds. The fastest way to reach me is email, and I usually reply within 24 hours. For freelance work I am reachable on weekday mornings and evenings, and most of the day on weekends and holidays (IST).