Backend & Distributed
Systems Engineer
|
01 / Skills
Things I use.
Languages
Backend
Cloud / Infra
AI Systems
Databases
Frameworks
02 / Experience
Things I've done.
Founding Engineer
Built and maintained backend services for three legal-tech products using Python and REST APIs, automated testing and deployment with GitHub Actions and AWS, owned a product end-to-end from PRD to deployment, and reduced monthly infrastructure costs by 10%.
Teaching Assistant
Taught core Java, Object-Oriented Programming (OOP), and basic Data Structures and Algorithms (DSA) to a summer cohort of 45+ students, leading to a 90%+ pass rate on the WWP Computer Programming district exam.
AI Fellow
Contributed primarily to Natural Language Human Feedback (NLHF) workflows by evaluating model responses, identifying quality gaps, and producing structured feedback for AI system improvement under strict quality and confidentiality requirements.
Laboratory Technician
Prepared sample batches for chemical analysis through preservation, spiking, reagent addition, digestion, and filtration, then documented sample preparation using Microsoft Excel scripts, LIMS (Laboratory Information Management Systems), and Clinisys software.
Education
Rutgers University, New Brunswick
Bachelor's in Computer Science · Minor in Economics · May 2027
03 / Projects
Things I've built.
Infrastructure-first projects emphasizing performance, scalability, and clean abstractions.
AutoTA
Deterministic automated grader that ingests 200+ Google Classroom submissions through the REST API, runs JUnit tests, and generates initial grades while detecting missing assignments.
challenge
Containerizing JUnit execution and building CI checks that keep grading safe, repeatable, and reliable across every commit.
JobMatch
Job ingestion and matching system that aggregates postings from 5+ job boards, stores raw data in S3 and normalized records in PostgreSQL, then ranks roles by career alignment.
challenge
Making job search repeatable by combining resilient ingestion, normalized storage, weighted NumPy ranking, and reliable test coverage for new sources.
AI-powered skin condition analysis app built at the MTC Hackathon. Users upload a photo to receive instant analysis across 8 condition classes, with confidence scores, severity ratings, personalized advice, and a DeepSeek-powered diagnosis chat interface.
challenge
Significant class imbalance in the SCIN dataset required careful augmentation and stratified splits. Calibrating the AI chat to be medically responsible without being overly restrictive was a real balancing act.
This Portfolio
The project you're viewing right now! A dark-mode developer portfolio built with Next.js 16 App Router, Framer Motion animations, Tailwind CSS v4, and deployed on Vercel.