Hello, I'm
Hello, I'm
AI Research Scientist working on interpretability, evaluation, and alignment of language models, with papers at ICLR 2026 (Oral) and COLM 2026. M.S. in Applied Data Science from USC, advised by Prof. Swabha Swayamdipta; Fulbright Scholar. Co-Founder & COO of Turon AI, building research-depth, production-grade B2B AI solutions for enterprises: computer vision systems for QA/QC and inspections and multimodal agentic systems for context ingestion and automation.
Understanding what language models actually know, building rigorous methods to measure it, and developing robust alignment solutions.
ICLR 2026 (Oral) · COLM 2026 · USC NLP
Research-depth, production-grade AI systems for manufacturing, automotive, and construction enterprises.
Turon AI · Retrocausal · Paramount
Applied AI Engineering roles
Strategic Partnerships & Customer Success: applied AI R&D and production engineering for manufacturing, automotive, and construction clients. We build research-depth, production-grade B2B AI solutions for enterprises.
Keywords
Technical Leadership & Solution Architecture: applied AI R&D and production engineering for manufacturing, automotive, and construction clients.
Keywords
Entertainment Analytics & AI: content performance forecasting and GenAI explainability for a major media organization.
Keywords
Computer Vision & Manufacturing Intelligence: precision assembly verification and ergonomics assessment for Fortune 500 manufacturing clients.
Keywords
Edge AI & Recommender Systems: model optimization for edge inference and a cloud-based recommendation engine.
Keywords
Research and Development roles
AI interpretability, alignment, reasoning, and evaluation research advised by Prof. Swabha Swayamdipta, mentored by Dr. Gregory Yauney, Matthew Finlayson, and Sayan Ghosh.
Keywords
Research & applied engineering case studies
Consensus Graphs (ConGrs): a DAG-based data structure that synthesizes epistemic signals across multiple sampled LM responses into a single, more reliable response. Accepted to COLM 2026.
A meta-evaluation measure for LLM micro-benchmarking that reveals when small evaluation subsets fail to reliably preserve model rankings from full benchmarks. ICLR 2026 Oral.
Real-time, camera-based assembly verification deployed to Fortune 500 manufacturing lines, achieving 98% detection accuracy while cutting manual annotation overhead by 80% through synthetic data generation.
A computer-vision ergonomics risk assessment module for continuous, objective monitoring of workplace safety and regulatory compliance on manufacturing floors.
Portfolio in progress. I'm actively open-sourcing projects to GitHub and adding new case studies here over time.
Peer-reviewed papers & preprints
S. Ghosh, S.S. Warraich, D. Tarsadiya, G. Yauney, S. Swayamdipta2025 – 2026Introduced Consensus Graphs (ConGrs), a flexible DAG-based data structure that captures shared information and semantic variation across multiple sampled LM responses to the same prompt, enabling task-dependent decoding into a single, more reliable response.
Keywords
Introduced a meta-evaluation measure for LLM micro-benchmarking that quantifies how reliably a micro-benchmark can rank two models as a function of their performance gap on the full benchmark, revealing when small evaluation subsets fail to preserve model rankings.
Keywords
Academic background & training
Graduate coursework and research in applied machine learning, natural language processing, and statistical methods at the USC Viterbi School of Engineering, including research on LLM interpretability and evaluation under Prof. Swabha Swayamdipta.
Key Courses
Technical stack across applied AI and research
Spoken and written proficiency
* DELF A2 certified
Student advocacy at USC
Viterbi Graduate Student Association (VGSA) · Los Angeles, CA 🇺🇸
Viterbi Graduate Student Association (VGSA) · Los Angeles, CA 🇺🇸
Course instruction at USC Viterbi School of Engineering
University of Southern California, Viterbi School of Engineering · Los Angeles, CA 🇺🇸
Writing on AI infrastructure, governance, safety, and building in applied AI
Recommended reading
Daniel Kahneman
A foundational read on how we think and decide: the interplay between fast, intuitive judgment and slow, deliberate reasoning.
Yuval Noah Harari
A sweeping account of how shared myths and cooperation shaped human civilization. Essential context for thinking about intelligence, natural or artificial.
Jordan B. Peterson
A thoughtful companion on discipline, meaning, and navigating complexity. Useful principles for both research and life.
Scholarships & honors
USC Viterbi School of Engineering
Awarded the USC Andrew J. Viterbi Scholarship in support of graduate studies in the M.S. Applied Data Science program.
U.S. Department of State
Selected as a Fulbright Scholar (U.S. Department of State, 2023 cohort), a flagship international educational exchange program, to pursue graduate studies at USC.
Let's work together
Phone
+1 213 681-6140
Schedule
Book a Call
X / Twitter
X / Twitter
GitHub
GitHub
Scholar
Google Scholar
Résumé
Download CV
Send a message
Send me a message and I'll respond as soon as possible.
