Pratham Shah

Resume

Software engineer and 2026 B.Tech Computer Science graduate from MIT-ADT University, Pune. Built Next.js, TypeScript and FastAPI features as a full-stack intern at NeuraMach AI Studios, and ships personal projects end to end, including a Chrome extension with 1,270+ weekly users and a serverless video platform on AWS.

Pune, India. Two one-page versions: one for software development roles, one for cloud roles.

Experience

Associate Cloud Reliability Engineer (Intern), Searce Inc.

Apr 2026 – Aug 2026

Pune, India

  • Analyzed Google Cloud billing data and resource utilization for client workloads to find idle and oversized resources.
  • Prepared rightsizing and cost-optimization recommendations, weighing savings against performance and reliability.
  • Tracked usage trends and cost allocation across GCP projects to support cost governance.

AI/ML Full-Stack Engineering Intern (QA focus), NeuraMach AI Studios

Jan 2026 – Apr 2026

Pune, India

  • Built responsive UIs in Next.js and TypeScript, using server-side rendering and static generation to improve load times and SEO.
  • Integrated FastAPI REST endpoints into the frontend and managed asynchronous data with React Query and Axios.
  • Used AI coding tools (GitHub Copilot, Cursor) to speed up development and tighten code reliability, alongside QA work on the product.

Frontend Developer (freelance), Abhinandan Mountreea

Jul 2025 – Nov 2025

Remote

  • Built the marketing site for a residential villa project in React, Vite and GSAP, from client requirements to deployment.
  • Served imagery through Cloudinary with WebP conversion, caching and lazy loading to cut page weight on mobile.

Projects

InsightVault: Question answering over uploaded PDFs

github.com/pshah-lab/insightVault
  • Built a Next.js app that chunks PDF text, stores embeddings in Supabase Postgres (pgvector, HNSW index) and answers questions from the top-matching passages through an LLM API.
  • Added per-IP rate limits, input length caps, PDF magic-byte checks and an embedding-provider fallback (Hugging Face, then Gemini, then OpenAI).

StreamVault: Invite-only video streaming on AWS

github.com/pshah-lab/StreamVault
  • Defined Cognito (hosted UI with PKCE), CloudFront signed cookies, a private S3 origin, Lambda, API Gateway and Secrets Manager in one AWS CDK stack.
  • Wrote the sign-in Lambda in TypeScript (HMAC-signed PKCE state, ID-token verification) with unit tests, and a FastAPI + DynamoDB service that saves watch progress behind Cognito JWT checks.
  • Scripted an FFmpeg pipeline that encodes uploads to HLS on an EC2 worker and syncs the output to S3.