Open to opportunities

Yagnesh Veeraraghavan

@yagneshveeraraghavan

Message

AI researcher and applied ML engineer building reproducible LLM experimentation pipelines for measurable breakthroughs.

United States

Message

What I'm looking for

I’m looking to join a research- and engineering-minded team where I can build reliable ML systems, run disciplined LLM experiments, and turn findings into shipped tools and product improvements.

I’m an AI Researcher who co-led authorship of a paper accepted to EACL SRW. I build end-to-end, reproducible LLM experimentation pipelines—covering data generation, fine-tuning, and evaluation—using Python, PyTorch, Unsloth, and Hugging Face.

I’ve extended experimentation to systematically test how 5+ training and data hyperparameters affect subliminal learning, and I designed token-isolation studies to isolate potential causes. My work disproved certain single-mechanism explanations and showed subliminal effects emerge from complex interactions.

I also bring hands-on software engineering experience, including mobile development with Expo/React Native, TypeScript, and Firebase. Alongside my research, I’ve built projects like a distributed key-value store (replication and causal consistency) and a multi-threaded HTTP server in C, and I value clarity, reproducibility, and fast iteration with teams.

Experience

Work history, roles, and key accomplishments

AI Researcher

Algoverse

Jul 2025 - Jan 2026 (6 months)

Co-led a research paper accepted to EACL SRW and built an end-to-end, reproducible LLM experimentation pipeline (data generation → fine-tuning → evaluation) using Python and PyTorch. Ran controlled experiments on subliminal learning by varying 5+ training/data hyperparameters and isolating training-data tokens to test competing explanations.

Python PyTorch Hugging Face UnSloth Experiment Design Hyperparameter Tuning Data Generation Model Evaluation

Mobile Developer

Blueprint Sea Otter Savvy

Sep 2023 - Jun 2024 (9 months)

Co-developed a nonprofit mobile application using Expo and React Native with TypeScript and Firebase, including an image upload feature that writes directly to Firebase. Contributed a picture creation flow with previews and agreement steps and supported mobile features for a real-time interactive sports finder experience.

Expo React Native TypeScript Firebase Mobile App Development