Curriculum vitae
Education
- Ph.D in Computer Science, University of Illinois Chicago, 2027 (expected)
- Advisor: Prof. Aadirupa Saha
- GPA: 3.87 — transferred from MS to PhD track based on research progress
- Research: LLM evaluation (AISTATS 2026), combinatorial bandits (NeurIPS 2025 OPT Workshop)
- B.Tech in Computer Science, Manipal University Jaipur (2019)
Work experience
- August 2022 - Present: PhD Student
- University of Illinois at Chicago (transferred from MS to PhD track)
- Location: Chicago
- Research collaboration with Adobe Research (Dr. Branislav Kveton) on LLM-as-a-Judge evaluation under a fixed budget — accepted at AISTATS 2026.
- Other research: combinatorial multi-armed bandits with monotonic aggregation (NeurIPS 2025 OPT Workshop); value smoothing via latent embeddings for RL; metareasoning for MCTS.
- Supervisor: Prof. Aadirupa Saha
- 2022 & 2023: Data Science Intern
- Dell Technologies — Austin, Texas
- Summer 2023 (Supervisor: Will Wilson): Applied prompt engineering and ChatGPT to summarize customer support conversations, reducing time-to-escalation for senior agents.
- Summer 2022 (Supervisor: Ashutosh Singh): Built an embedding-based search system to surface answers from Dell’s knowledge base; deployed to the live support website.
- November 2019 - March 2021: Deep Learning Engineer
- Sensight Labs (Friendly Brands now)
- Location: Bangalore
- Built an intelligent shopping cart system to eliminate checkout queues; developed computer vision pipelines to identify grocery products. Projects: cart detection, grocery web scraping, automated data analysis, colour-based item selection, visual text-based search engine, video-to-image cropping tool.
- Supervisor: Prashant Maurice
- August 2019 - November 2019: Machine Learning Engineer
- IceCream Labs
- Location: Bangalore
- Built a zero-shot brand identification model (front-image only); 97% accuracy across 250 brands with no additional training required — infinitely extensible to new brands.
- Supervisor: Bharat
- January 2019 - April 2019: Analytics Intern
- Spinnaker Analytics
- Location: Mumbai
- Delivered predictive analytics models for client engagements; built a time-series forecasting system to predict Boston Children’s Museum footfall and a return-prediction model for e-commerce orders.
- Supervisor: Ashwin Meshram
- June 2018 - August 2018: Machine Learning Intern
- IceCream Labs
- Location: Bangalore
- Developed CNN-based pipelines (Keras, OpenCV) to cluster visually similar products for e-commerce recommendations; built nutrition-table OCR pipelines and a synthetic data generation pipeline. Projects: Visual Search, Fashion Design Overlay, Variable MIL, Nutrition Table Mapping.
- Supervisor: Prashant Maurice
- June 2017 - August 2017: Automation Intern
- CSS Corp (Movate)
- Location: Chennai
- Built Python automation tools for internal teams (VLookup automation, workflow scripts); contributed plugins to open-source tools including NVDA.
- Supervisor: Kiran Marri
Skills
Languages & Tools: Python, PyTorch, TensorFlow, NumPy, scikit-learn, OpenCV, Pandas, Git, AWS, LaTeX, HuggingFace Transformers, OpenAI / Anthropic APIs, prompt engineering.
Research Areas: Multi-Armed Bandits, Online Learning, Reinforcement Learning, Convex Optimization, Theoretical Machine Learning, LLM Evaluation, Deep Learning, NLP, Computer Vision.
Awards
- NeurIPS 2025 Conference Travel Grant — competitive travel grant awarded to attend NeurIPS 2025 in San Diego, CA (Dec 2025).
- MS → PhD Track Transfer — transferred to the PhD program at UIC based on research progress.
- Microsoft Imagine Cup — Team placed in the top 8 nationally during undergrad.
- IISc–IBM Pravega Hackathon — Awarded 2nd place.
Talks & Presentations
- LLM-as-Judge on a Budget — to be presented at AISTATS 2026 (Morocco, 2026).
- Efficient Algorithms for Combinatorial-Bandits with Monotonicity — NeurIPS 2025 OPT Workshop, San Diego, CA, Dec 2025 (poster).
Academic Service
- Reviewer, International Conference on Machine Learning (ICML) 2026.
Publications
See the full Publications page for papers with abstracts, links, and code.