CV
Curriculum vitae
General Information
| Full Name | Raj Sanjay Shah |
| rajsanjayshah@gmail.com | |
| Website | https://raj-sanjay-shah.github.io |
| Research Areas | LLM safety, cognitive modeling, AI for mental health |
| Languages | English (native), Hindi (native), Gujarati (native), Marathi (professional), Sanskrit (intermediate) |
Education
-
2023 - Present PhD in Computer Science
Georgia Institute of Technology · Interactive Computing - Advisor - Prof. Sashank Varma
- Focus on LLM safety, developmental alignment, and AI copilots for well-being.
-
2024 - 2025 Visiting Student Researcher
Stanford University · SALT Lab - Host - Prof. Diyi Yang
- Joint projects on counselor coaching systems and unlearning brittleness.
-
2021 - 2023 MS in Computer Science (Machine Learning)
Georgia Institute of Technology - GPA 4.0 / 4.0
- Coursework in ML, NLP, and human-centered computing.
-
2017 - 2021 BE in Computer Science + Minor in Data Science
BITS Pilani, India - CGPA 8.54 (Major), 9.8 (Minor)
- Graduated with institute best-paper honors in Information Retrieval.
Experience
-
Summer 2025 Applied Scientist Intern II
Amazon Science · Rufus - Built clarification-driven dialogue summarization that detects missing info and auto-asks targeted questions.
- Reduced omission errors >25% across multi-domain transcripts.
-
Summer 2024 Applied Scientist Intern II
Amazon Science · HealthScribe - Evaluated LLM note generation for behavioral therapy progress notes.
- Delivered quality metrics now embedded in AWS HealthScribe workflows.
-
Summer 2023 Sr. Research Software Engineer (AI)
IBM Research · Human-Centered AI - Ran user studies on calibrated human-AI collaboration for counseling support.
- Shipped full-stack tooling for real-time MI coaching experiments.
-
Summer 2022 Software Developer · ML Intern
Epic Systems - Modeled clinician workflows to surface relevant patient context in dashboards.
-
Jan 2021 - Jun 2021 Software Development Intern
JPMorgan Chase & Co. - Automated regression suites for the merchant onboarding platform.
-
Summer 2020 Research Intern
Samsung R&D Institute India - Built a bi-LSTM VAE to generate melodies/chords for the ringtone UX program.
Open Source & Community Projects
-
2023 - Present BabyLM Challenge + Cognitive Batteries
- Co-organizer of BabyLM workshops, released staged corpora, evaluation tooling, and dev-leaderboards.
-
2022 - Present Counseling Copilot Suite
- Open datasets + MI annotation guidelines powering LLM-based counselor practice sandboxes.
Honors and Awards
-
2023 - Georgia Tech President's Fellowship
- Marshall D. Williamson Fellowship (Honors Program)
-
2022 - Donald V. Jackson Fellowship (Honors Program)
- LRN Foundation Student Travel Grant (EMNLP)
-
2021 - BITS Pilani Best Paper Award - Information Retrieval
Academic Interests
-
LLM Safety & Control
- Unlearning, watermarking, continual learning
- Stress-testing model behavior in applied settings
-
Cognitive Alignment of PLMs
- Developmentally staged corpora
- Linking hypotheses between psych signatures and model activations
-
AI for Mental Health Support
- Counselor copilots and reflective practice
- Evaluation protocols for therapy documentation
Other Interests
- Hobbies: Long-form storytelling, running Atlanta’s beltline trails, mentoring undergraduate researchers.