AI/ML researcher and engineer completing my MS in Artificial Intelligence at Long Island University, focused on LLM fine-tuning, agentic systems, and the evaluation work that separates a model that looks good from one that actually works.
New York, USA
I'm Rashmi — an AI/ML researcher in New York, finishing my MS at Long Island University. I build LLM agents and fine-tune models, but honestly, I care more about whether they actually work than whether they look good in a slide.
Right now I'm at Anote working on evaluation and agentic systems, with a paper under review at AAAI 2027. Outside of that: hackathons, a YouTube channel for my side projects, and a running list of new restaurants to try.
The stack behind the research and the products.
Building Panacea's AI-assisted coding agent — orchestrator, crew, task, and tool-use components.
LoRA & adapters, synthetic SFT data, and the eval work behind an AAAI 2027 submission.
PyTorch, transformers, and multimodal CV + NLP systems shipped to production.
DP-SGD within fine-tuning pipelines, quantifying privacy/utility tradeoffs at scale.
Research, publications, and things I've built.
Moments from hackathons, forums, and open-source work. Send me your photos and I'll drop them right in.
Built an AI-powered multimodal tutor exploring Gemini 3 Pro's advanced reasoning and native multimodality in Google AI Studio — combining text, visuals, and context to create interactive, intelligent STEM coaching.
Built in 4 hours: an AI system that compares two consecutive pharma earnings call transcripts to automatically surface red flags — pipeline omissions, milestone delays, guidance cuts, sentiment shifts, competitive risks, and disclosure changes. No more reading 40-page transcripts line by line.
Built a voice assistant designed to make financial fitness more accessible — hands-free budgeting, spending insights, and financial coaching, built end to end in a single hackathon sprint.
Represented Anote at CMU's Robotics Innovation Center, engaging with Army program leads, defense primes, and AI/robotics companies on synthetic data and explainable AI.
Research contributions to NumFOCUS open-source AI projects.
Peer-reviewed computer vision research achieving 94% classification accuracy on a 1000+ image dataset.
Notes, newsletters, and videos on AI/ML and what I'm learning.
I write about AI/ML, agentic systems, and what I'm learning as a researcher.
Newsletter-style deep dives on AI/ML research and what I'm building.
Videos documenting personal AI/ML projects, builds, and experiments.
A few things I love outside of work. Photos coming soon.
I love exploring new restaurants and cuisines — good food is one of my favorite ways to unwind and explore a city.
Exploring new places recharges me and keeps me curious — from city skylines to quiet beaches.
Training keeps me disciplined and clear-headed — it's the same focus I try to bring to research and engineering.
Have a question or want to work together? Send me a message.