Hi, I'm Adhya Dagar.
I work at the intersection of research and product, designing evaluation systems that help teams understand and improve AI behavior.
Much of my work involves translating between researchers and domain experts — turning real-world workflows and human judgment into signals that models can learn from.
Work
Adobe · Firefly
Product Manager, Model Evaluations
Pareto.AI
Product Manager, Research & Evaluations
BitClass
Product Manager
Education
Fellowships
Felicis Ventures · AI Fellows
Felicis AI Fellow
Corporate Startup Lab · CMU Swartz Center
AI Strategy and Research Fellow
Lab
The Three Selves Eval Framework (Draft)
Mar 2026Behavioral framework for measuring LLM identity stability under adversarial pressure.
Research
3D Product Placement in Ad Creatives
May 2024Tools for generating hyper-realistic backgrounds for 3D renders with SGK, reducing production costs and time.
Generative AI
ProjectArena — Evaluation Suite for Project Management Agents
April 2024Evaluation framework for autonomous agents in project management using Dockerized environments (Plane, Rocket.Chat).
Evaluation
RAG for Academic Question-Answering
Feb 2024Retrieval-augmented generation system for CMU domain QA. 293-entry test dataset, multiple model experiments including Mistral and DeBERTa.
NLP
Collaborative Agents for QA Generation
November 2023Three-agent framework for cancer-domain Question Answer synthesis. 800+ entry dataset, 40%+ Turing Test achievement.
Research
Corporate Startup Lab — Shell
June 2023Case study on Shell's approach to AI adoption and enterprise innovation through CXO interviews.
Case Study
Driving Organizational Transformation — Enterprise AI
May 2023Research on Fortune 500 AI adoption through interviews, state-of-the-art reports, and organizational readiness surveys.
AI Strategy
Clay/Play
Pottery Wheel
Shape clay, pick a glaze, fire it in a kiln.
Moments
Raw Cuts
a good city deserves a love song