Keshavan Seshadri Email & Phone Number

Newark, New Jersey, Us

Developed and deployed Table-Augmented Generation (TAG) to enable natural language querying of complex insurance and financial data from Excel/CSV files within the chatbot.

Built an innovative grant discovery and application tool from the ground up as a part of Cornell Tech's startup studio course using Retrieval-Augmented Generation (RAG), GPT-4, BERT, Docker, AWS App Runner, RDS, Cognito, EC2, GitHub Actions CI/CD, Flask, and React. Designed in Figma and developed the MVP (Minimum Viable Product) from 0 to 1, integrating public datasets to capture a $34 billion market opportunity with targeted early revenue projections of $85 million. Gained significant traction with a growing user base and potential early customers, including nonprofits, entrepreneurs, and SMBs, and interviewed with accelerators. Secured a spot in the Cornell Johnson Summer Startup Accelerator.

Dublin, Dublin, Ie

Implemented new features and version updates for test frameworks Playwright, Puppeteer, and Cypress on remote machines keeping the system up to date. Contributed to the framework maintenance and debugging of the company's "Automate" product, ensuring its uninterrupted performance and minimizing software defects for enhanced product stability.

Bengaluru, Karnataka, In

Designed and developed a GUI-based platform for training and inferring ResNet, Inception, and Deep Speech models, leveraging a full-stack solution with Docker, Kubernetes (k8s), and Django. Engineered an AI platform enabling users to easily upload news videos, receive timestamps of flagged Violent and NSFW content, and obtain censored versions, ensuring compliance and promoting a safer viewing experience.

Kingston, On, Ca

Worked under Dr. Farnaz Heidar-Zadeh on physics-based Machine Learning models to predict quantum-mechanical observables such as energies and forces with the QM17 dataset.

Hyderabad , Telangana, In

Conducted tutorials and evaluated over 200 students’ performance for Science-1 and Science-2 courses.Motivated and managed over 20 students in projects outside the coursework such as building a Tcl Script library to analyze large molecular dynamics trajectories

Extended the open-source 3DMol.js library to support a clicker-based active learning environment and a Flask server to allow interactive viewing and querying of 3D molecules.Established seamless real-time synchronization of the 3D molecular state across 1000s of users with minimal latency by sending delta state-updates to the server using SocketIO. 𝗞𝗲𝘀𝗵𝗮𝘃𝗮𝗻 𝗦𝗲𝘀𝗵𝗮𝗱𝗿𝗶, Peng Liu, David Ryan Koes; The 3Dmol.js learning environment: a classroom response system for 3D chemical structures. Journal of Chemical Education, 2020 97 (10), 3872-3876DOI: 10.1021/acs.jchemed.0c00579

Hyderabad , Telangana, In

Taught Computational Thinking, applications, and Problem-Solving using game-based fun activities to over 240 students of classes 7-10. The students widely appreciated my style of teaching and I was called "Future Maker" and "Feedback Master" by the students.

Hyderabad, Andhra Pradesh, In

Developed a game-based learning tool used by millions of students to explain Data Structures.Organized, authored, and delivered workshops to 200+ industry enthusiasts and entrants on basics of programming, open-source development, and web development.

Master'S Degree, Computer Science

Master Of Engineering - Meng, Computer Science

B.Tech In Computer Science And Ms By Research In Computational Natural Sciences, Computer Science

Education record