Responsible for the development of the entire stack for LiveDuel, which is a decentralized sports prediction market backed by Lead and Techstars. I worked closely with my CEO, initially as the only member of the development team, to design the architecture for the MVP. It grew to raise 250K+ in pre-seed funding from Avalanche.• Built the frontend(NextJS), backend, services, sdk(ethers) and other internal tools for the company.• Customized smart contracts(solidity) to fit the company's needs. Added layers of security and control to the contracts.• Designed the architecture to accommodate a large volume of trading and real-time price updates on charts.• Dockerized the stack for rapid development cycles and easy onboarding. New teammates were able to start working on tickets within a couple of days of joining.• Managed the deployment using AWS EKS & Kubernetes. Analyzed usage through load testing and brought down costs by 70% compared to earlier stages.

Worked on reinforcement learning research projects and was responsible for managing the Flatland challenge on the platform. Also responsible for another product called Blitz Weekly, where we released unique challenges every week. Since it was a very tiny company, worked closely with my CEO and CTO to take projects from ideas to being practically implemented.• Built modules and designed the architecture for the challenges that accommodated anywhere from 300 to 1500 users per challenge and around 5000 submissions on average for each challenge• Created evaluation containers that would assess users’ submissions. Submission flow was automated with GitLab CI/CD. This improvement saw a massive increase in submissions (around 10-15x) as it involved very little effort from the end user.• Gained extensive experience working with AWS and managing all the resources and infrastructure required for deploying entire applications and achieving end-to-end automation.

Hippo Video was an AI driven video platform that provided tools to create videos for sales and marketing. It was a relatively bigger company than the other ones I worked at, I worked with the ML team which was under-staffed at the time and setup their process for deploying ML models to be integrated with their platform.• Designed the backend and utilized AWS resources to dynamically scale based on active jobs and users using the ML models. Containerized jobs/tasks run on demand and can accommodate any number of users.

Verwin was a job portal forindividuals who needed assistance during theirjob search. The company was being set up during my internship.• Primary work on the backend involved setting up routes for access to MongoDB, configuring authentication with Passport.js and integrating Elasticsearch for a better searching experience based on titles, keywords, and tags.• Worked on integrating the backend with the React web app.• Got exposed to the thought process behind designing an architecture that scales. The company grew to accommodate around 17,000 users before eventually shutting down operations.