PennOS
Unix-like OS in C with priority scheduler, system calls, and thread-based process model. Features FAT-style file system and shell supporting POSIX-like operations, pipelines, and I/O redirection.
Building privacy-preserving AI systems.
Master's student at Penn Engineering.
I'm a Master's student at Penn Engineering passionate about building at the intersection of AI, systems, and real-world impact. My research background in privacy-preserving ML and federated learning gives me a unique lens. I care deeply about building AI that's not just powerful, but trustworthy.
My experience spans data analytics, full-stack and production-grade software development, and applied ML research. From shipping real systems at JPMorganChase to publishing federated learning work at IEEE conferences. I thrive in fast-paced environments, love tackling hard problems and am at my best when moving quickly from idea to working product.
Right now, I'm actively looking for Summer 2026 internships in AI/ML engineering, applied research, or full-stack roles. I'm also genuinely excited to connect with founders and builders whether you're early-stage or just exploring an idea, I'd love to collaborate on something ambitious.
Won 1st place in internal hackathon building an NL-to-SQL generator using LLMs and prompt engineering. Leading AWS cloud migration and API integrations with new vendors.
Built report scheduler dashboard and enhanced reporting module. Conducted LLM onboarding sessions. Closed 75+ Jira stories and handled 6 independent production releases.
Built predictive models for contract risk using Salesforce data. Designed interactive CRM Analytics dashboards. Trained on 70+ SAQL use cases.
Researched privacy-preserving recommendation systems using federated learning. Built graph neural networks with attention mechanisms. Published findings at IEEE conference.
Automated AWS resource tracking reports eliminating 100% unnecessary usage. Built REST APIs with FastAPI and Redis for real-time crypto prices with Slack bot integration.
Master's in Computer and Information Sciences
CGPA: 3.7 Philadelphia, USAB.Tech in Computer Engineering
CGPA: 3.94 Mumbai, IndiaPrivacy-preserving federated learning system for medical literature Q&A. Enables hospitals to collaboratively answer queries using Flower, Qdrant, and LangChain without sharing sensitive patient data. HIPAA-compliant with 3-5 sec response times.
Privacy-preserving recommendation system using FedScope architecture with FedME and FedEM algorithms. Solves cold-start problem with meta-learning.
Unix-like OS in C with priority scheduler, system calls, and thread-based process model. Features FAT-style file system and shell supporting POSIX-like operations, pipelines, and I/O redirection.
Comprehensive virtual research assistant with collaborative workspaces, in-browser PDF viewer, annotator, and AI-powered summarization. Top 10 at KJSCE Hack 7.
Chrome extension that blocks hateful and offensive comments on YouTube and Twitter using NLP-based toxicity detection with TensorFlow and FastAPI.
Personalized banking system with ML models for credit risk scoring, churn prediction, and tailored product recommendations using Next.js and Firebase.
Futuristic sustainable investment platform promoting ESG investments with portfolio generation and advisory for both companies and investors.
Disease prediction system for breast and cervical cancer detection using Logistic Regression and Random Forests. Built at SheHacks hackathon.
Investigates vulnerabilities in Federated Learning protocols, exploring inference attacks and proposing mitigation strategies to enhance privacy and security.
Integrates Blockchain with FL using SMPC encryption, Committee Consensus Mechanism, smart contracts, and IPFS for secure personalized healthcare models.
Proposes Graph Federated Learning with attention mechanisms and meta-learning to address cold start scenarios while preserving user privacy.
Proposed FedScope Recommendation System (FRS) for next keyword prediction enabling personalized recommendations while preserving user privacy.
Built an NL-to-SQL generator using LLMs and prompt engineering
2024Developed SpaceVest, a sustainable investment platform promoting eco-friendly options
2023Mentored underprivileged students to build a solution for detecting paddy plant diseases
2024Led PR for the official student-run newspaper across 3 university campuses
2022-2023Provided basic reading and writing lessons to help with daily literacy needs
2021