Namaste. Hello, I'm Shivam.
Building responsible, steerable, and interpretable AI systems.
Ph.D. candidate in Computer Science working at the intersection of generative AI, multimodal foundation models, machine unlearning, and responsible model editing. My research explores how intelligent systems can be made more capable, interpretable, controllable, and accountable — alongside that, I've also co-founded a startup building generative video products.
About
About Me
I was born in Ranchi, when it was still part of Bihar, and my family's roots trace back to Nalanda — an ancient seat of learning whose long association with scholarship and inquiry has quietly shaped how I think about my own work. That thread runs from those early questions to where I am today: a Ph.D. candidate in Computer Science at West Virginia University, working on generative AI, responsible model editing, and multimodal intelligence.
My work moves between research questions and working systems — from neuron-level methods for unlearning and editing concepts in foundation models, to co-founding a startup that turns language into generated video. I like living in both places at once: the rigor of a research question and the discipline of shipping something real.
- Location
- Morgantown, WV, USA
- Institution
- West Virginia University
- Open to
- Research ScientistApplied Research ScientistMachine Learning Research ScientistGenerative AI EngineerMachine Learning EngineerAI Research EngineerPostdoctoral Researcher
Research
Generative AI, Responsibly
My research moves between generative AI, multimodal foundation models, and making those systems easier to understand, edit, and trust.
IEEE IJCB 2025
CURE: Centroid-guided Unsupervised Representation Erasure for Facial Recognition
Removes the influence of specific identities from facial recognition models through centroid-guided, unsupervised representation erasure — without full retraining.
Under Review, IEEE T-BIOM
NEST: Fine-Grained Concept Unlearning via Neuron-Level Targeted Editing
Targets and edits individual neurons to perform fine-grained concept unlearning in foundation models, aiming for precise removal with minimal collateral impact on unrelated capabilities.
Startup — Valoi
Valoi Generative Video System
An AI-powered text-to-video system that converts educational text into videos for dyslexic learners — combining LLM-driven reasoning and scene planning with diffusion-based generative video into a unified, multi-scene pipeline. Underlies two U.S. patent applications.
Career
Professional Experience
Where research and applied engineering have met in practice.
Co-Founder
Jan 2024 — PresentBuilt an AI-powered text-to-video tool that converts educational text into videos designed to support dyslexic learners.
- Designed advanced LLM-driven reasoning and scene-planning techniques alongside novel diffusion-based text-to-video methods.
- Integrated both into a unified generative video system underlying two U.S. patent applications.
- Worked across research, architecture, backend systems, inference, evaluation, optimization, and deployment.
Graduate Research & Teaching Assistant
Jan 2022 — PresentResearch on neuron-level and sparse-representation methods for concept unlearning and editing in foundation models.
- Developed neuron-level and sparse-representation methods for concept unlearning and editing in foundation models.
- Designed privacy-preserving representation-erasure algorithms for biometric recognition systems.
- Built and evaluated model-steering methods for CLIP, Stable Diffusion, and multimodal foundation models.
- Teaching Assistant for CS 111 (Data Structures), CS 474/574 (Responsible AI), and CS 480/481 (Capstone); led labs and mentored undergraduate student projects.
Research Intern
Jun 2024 — Aug 2024Worked on healthcare projects using generative AI, notably MRI tumor analysis using diffusion models.
- Explored MRI tumor analysis using diffusion-based generative models.
- Collaborated with hospitals and academic researchers on AI integration in medical research.
Toolkit
Technical Skills
Core tools and frameworks across research and production engineering.
Generative AI
ML Engineering
Responsible AI & Research
Infrastructure & Deployment
Languages & Frameworks
Academics
Education
Ph.D. Candidate in Computer Science
West Virginia University
- Research focus: generative AI, multimodal foundation models, machine unlearning, and responsible model editing.
Bachelor of Engineering in Computer Engineering
Savitribai Phule Pune University
Beyond the Lab
A Bit of My Journey
Conferences, travel, and a few other things that round out the research — the full story lives on the Journey page.
Get in Touch
Contact
Open to research collaborations, applied research and Generative AI / ML engineering roles, speaking or academic conversations, and discussions on responsible AI and machine unlearning. The fastest way to reach me is email.