Document 01

Résumé

Education, engineering roles, and the projects I keep coming back to. It is the same page I send out, set in the site’s type.

Shaurya Verma

Education

May 2028

Bachelor of Science (BS), Electrical Engineering and Computer Sciences

GPA: 3.7

Relevant coursework: Data Structures, Discrete Mathematics & Probability, Structure and Interpretation of Computer Programs.

Experience

April 2026 – Present

Software Engineer

Redwood City, CA, USA

Owned end-to-end development of a Next.js, Supabase, and PostgreSQL CNC operations platform used to manage 150+ orders across quoting, payments, production scheduling, and shop-floor execution.
Built deterministic CAD quoting workflows using geometric complexity, material, machine load, and required operations; benchmarked estimates against 50+ historical orders and added a Claude Haiku review layer grounded in similar parts and order history.
Automated G-code readiness, setup-sheet generation, per-part file handoffs to Raspberry Pi-connected machines, KPI telemetry, and daily production prioritization.
Added 2,000+ Vitest unit and Playwright end-to-end tests and implemented GitHub Actions CI/CD, role-based access controls, and audit logging, replacing days of manual regression work and preventing order-state failures.

March 2026 – May 2026

Founding Engineering Intern

Remote

Built React Native/web heart-health workflows and re-architected Ask Veevo, a FastAPI LLM assistant, with on-demand context retrieval and streamlined prompts, reducing response latency from 15–20 seconds to 3–5 seconds.

December 2025 – March 2026

Founding Engineer

Remote

Led full-stack development of an AI-assisted GCSE revision platform serving ~10,000 monthly learners, building personalized diagnostics, practice workflows, mastery tracking, and daily study plans.
Architected authentication, database architecture, payments, and usage-based access control while coordinating intern contributions, moving the platform from prototype to production.

June 2024 – August 2024

AI Developer Intern

Pleasanton, CA

Developed a PyTorch multi-task learning component for a RAG-based payer-response prediction system evaluated on 1,000+ healthcare claims; built structured-data preprocessing, feature construction, and training/evaluation pipelines.

Projects

Algorithmic puzzle solver and full-stack platform

Python · FastAPI · React/Next.js · search algorithms

Built and deployed a FastAPI/Next.js puzzle solver using push-level state search, flood-fill reachability, transposition caching, deadlock pruning, and verified move reconstruction, solving 99.12% of 381 boards within a 30-second limit.
Developed a portfolio of weighted A*, greedy and beam-search variants, macro successors, and learned ranking, with automated replay verification and benchmarking against published solutions and BFS/A* baselines.

Finalist, Intelligent Planet Hackathon

AI pest detection platform

Expo · GCP · Python · ML

Selected as one of 25 global finalist teams in the Intelligent Planet Hackathon for an acoustic ML system detecting early red palm weevil infestations.
Trained and deployed a spectrogram-based DenseNet on Vertex AI, achieving 92% accuracy on 10,000+ augmented samples generated from held-out source recordings; built an Expo mobile interface for field-data ingestion and detection.

Skills