I'm a Computer Science graduate student at Univeristy of California, Davis passionate about AI and Machine learning. My research focuses on deep learning, computer vision, CNNs, and transformer-based architectures. On the applied side, I build LLM-powered applications, RAG systems, and AI agents; designing intelligent workflows that work in production. I'm also expanding into data science, and across all my work I focus on building AI that is practical, scalable, and built for real-world impact.
University of California, Davis, California
University of California, Davis, California
TransferX, California
JVS Engineering, India
Chemistry Department · University of California, Davis, California
Electrical & Computer Engineering Department · University of California, Davis, California
University of California, Davis, California
Savitribai Phule Pune University, India
Savitribai Phule Pune College, India
Sardar Dastur Hormazdiar Junior College, India
Mount Carmel Convent High School, India
An offline-first Android travel assistant that runs entirely on-device, with no internet and no cloud. Uses LiteRT for on-device LLM chat and translation, ML Kit for real-time OCR, and a local vector RAG system over a knowledge pack to answer contextual travel questions anywhere in the world.
An offline-first disaster communication platform designed to maintain critical connectivity during network outages and emergency scenarios. It enables peer-to-peer messaging, local resource sharing, and decentralized SOS coordination without stable internet, with resilient synchronization once connectivity is restored.
An agentic materials copilot built to accelerate critical mineral discovery. Takes mission-level language, converts it into search constraints, and queries the Materials Project with resilient retrieval. A planner loop scores and filters candidates iteratively and delivers ranked material portfolios with uncertainty estimates and lab test queues, all streamed live.
A shared household management app that replaces the mess of group chats and spreadsheets. Supports Google sign-in, house onboarding, and admin/member roles, with a unified dashboard for expenses, chores, shopping lists, announcements, and a shared calendar, all synced in real time across housemates.
An AI-powered query tutor for ResilientDB's GraphQL API. Ingests GraphQL documentation into a RAG pipeline so users can ask natural language questions, get structured explanations, complexity estimates, and efficiency tips without digging through docs, served through a clean Nexus UI with Dockerized backend services.
An end-to-end revenue pipeline built on hotel booking data. Uses Prophet for demand forecasting, de-trended elasticity curves to model price sensitivity, constrained nonlinear optimization for pricing decisions, and multi-period dynamic programming to maximize revenue across booking windows, validated with A/B-style simulation.
A production-grade retail forecasting system that goes beyond naive baselines. Engineers causal features from promotions, pricing, and calendar effects, fits probabilistic quantile models for uncertainty-aware predictions, reconciles forecasts hierarchically across SKU, store, and geography, and monitors for distribution drift to stay reliable in production.
A customer analytics pipeline that segments users by recency, frequency, and monetary value to identify high-value cohorts and churn-risk clusters. Predictive retention models then target each segment with tailored interventions, replacing one-size-fits-all campaigns with ROI-driven retention strategy built directly from transactional history.
The National Science Foundation's Predictive Intelligence for Pandemic Prevention (PIPP) initiative and the Center for Pandemic Insight (CPI), in partnership with the CHPS Lab, are building the early warning infrastructure the world needs to get ahead of the next pandemic — before it reaches human populations.
Highly Pathogenic Avian Influenza (HPAI) is no longer a contained wildlife problem. Recent outbreaks have demonstrated sustained transmission from migratory birds to marine mammals, and every cross-species jump brings us closer to a strain capable of spreading between humans. Harbor seals sit at exactly this interface, they are one of our most valuable sentinel species for detecting the early signals of a zoonotic spillover event. The challenge is that meaningful surveillance requires identifying and tracking individual animals across time and space at a scale that manual field methods simply cannot reach.
SealID is a computer vision system built to solve that problem. By automating individual seal identification from drone footage and field photographs, it enables non-invasive, continuous, large-scale wildlife surveillance that feeds directly into CPI's pandemic early warning pipeline. Rather than waiting for a die-off to be reported, the system tracks population-level changes in behavior, body condition, and mortality in real time, turning wildlife monitoring from a reactive process into a predictive one.
Leading the computer vision development on this project, the focus has been on a full architectural upgrade of the identification pipeline. The previous system relied on a CNN-based model with manual face chipping, feature annotation, and per-seal retraining, steps that created significant bottlenecks at every stage. The rebuilt system replaces all of that. Face detection and pose-aligned chipping are now handled automatically by YOLOv8-Pose, and identity matching is performed by MegaDescriptor, a Vision Transformer, which encodes each seal face into a 2048-dimensional embedding and matches it against a FAISS vector database of 130+ individually catalogued harbour seals with 300 to 400 reference images per individual. The system achieves 96% face detection accuracy and 73% individual identification accuracy on unseen field data, and extends reliable recognition to side profiles, partial faces, and multi-individual frames that the previous system could not handle. Adding a new individual to the database no longer requires retraining, it requires dropping photos into a folder and running a single script.
A desktop GUI allows researchers to upload any photograph and immediately receive ranked identity matches with confidence scores, making model outputs interpretable and actionable without requiring any technical background. The broader goal is a system robust enough to handle real-world drone footage conditions, varying altitude, angle, lighting, and motion, so that pandemic surveillance can run continuously without human bottlenecks slowing it down. By enabling consistent longitudinal tracking of individual animals, SealID provides the behavioral and mortality data needed to detect anomalies before a spillover event reaches human populations.
Covers the full analytics workflow: framing business problems as data problems, selecting and evaluating models, and translating findings into executive-level strategy. Bridges technical data science with real-world business impact.
Learned core machine learning concepts through hands-on projects involving supervised and unsupervised learning, decision trees, ensemble methods, recommender systems, and practical model development using Python, TensorFlow, and scikit-learn.
Completed an advanced specialization covering deep neural networks, hyperparameter tuning, optimization techniques, convolutional neural networks (CNNs), sequence models, TensorFlow, and best practices for designing and deploying deep learning systems.
Developed practical prompting skills for working effectively with modern AI models. Learned how to craft better prompts for research, writing, brainstorming, image generation, and building simple AI-powered applications while improving the quality and reliability of AI responses.
Built a strong foundation in generative AI by exploring transformer architectures, foundation models, fine-tuning techniques, responsible AI practices, and deployment strategies for developing production-ready LLM applications on AWS.
Designed and deployed interactive GenAI applications using Streamlit and Snowflake Cortex AI. Applied prompt engineering, Retrieval-Augmented Generation (RAG), and rapid prototyping techniques to build AI-powered applications from concept to working prototype.
End-to-end SQL training from core querying through joins, aggregations, subqueries, and analytical workflows. Built practical database skills for extracting, transforming, and analyzing data with hands-on exercises.
Outside of work, you'll usually find me by the ocean with a camera, chasing skies that never look the same.
If you have an opportunity, a project, or something you think is meant for me, I’d love to hear it. Let’s chat over coffee, or just a well-timed email.