Merit-based scholarship awarded to the top 1% of undergraduates at UC Berkeley.
Top 161 students in the nation, awarded by President Joe Biden.
I'm Manav, and I'm a Regents' Scholar studying Electrical Engineering and Computer Science at Berkeley. I'm currently seeking 2026 internships, and most of my experience has been in infrastructure/systems, machine learning, and full-stack development.
Outside of school, you'll find me cooking, playing soccer, or building creative ML projects with friends in LaunchPad!
I love learning anything and everything. Feel free to reach out anytime!
Automated Scaling Team
I built an end-to-end predictive host count tool which improved our customers' performance efficiency by 36% on average.
Alignment Team
I developed a reinforcement learning with human feedback (RLHF) pipeline to ensure our LLM did not give medical advice nor false information.
University Team
I built a full-stack data analytics dashboard to track institutional donation and outreach trends.
NASA
I conducted software-side tests on the personnel door gaskets for NASA's Artemis II Rocket. I also worked with the R&D team to model crystal behavior during nuclear fission.
I sequenced and processed 20M+ genetic sequences to improve taxonomic accuracy for Liriope muscari. Our results were published in BOLDSystems.
We build creative Machine Learning projects! We also host ML paper reading groups and workshops. As Vice-President, I organize our external client engagements and ensure we're fiscally responsible.
We solve real-world business and technology problems for Fortune 500 Tech companies. I currently serve as a development mentor where I help new members build their technical and presentation skills.
I lead biweekly tutoring sessions for 5+ students in CS 61A, teaching core concepts and guiding them through practice problems. As a Senior Mentor, I also coach fellow mentors on effective teaching strategies/pedagogy.
Designed a social app for discovering and sharing local finds with friends through a QR code-based friend system, featuring personalized recommendations and interactive maps with real-time geolocation tagging.
Built an NFL/NBA player proposition predictor with 75% inference accuracy using a modified DeepGBM architecture.
Created a secure paper-trading platform using Flask with real-time stock market data integration, enabling users to practice trading without financial risk through a scalable NoSQL database system.
Developed a Web3-powered platform for college students to securely buy, sell, and trade textbooks via NFT-based campus verification, IPFS storage, and trustless stablecoin escrow smart contracts on Polygon.
Developed a complete recreation of the classic tower defense game with custom graphics, game mechanics, and interactive gameplay elements.
Built a bigram language model using PyTorch that impersonates Yoda's speech patterns.
Developed an algorithm that identifies people based on analysis of their Short-Tandem-Repeats (STRs), implementing DNA profiling principles used in forensic analysis with CSV database matching.
Developed a text readability analyzer implementing the Coleman-Liau index to automatically determine U.S. grade-level comprehension from written content.