Anay Gupta
Software Engineer
Software Engineer with experience in Python, TypeScript, and C, focused on backend development, RESTful APIs, and system automation. Hands-on experience with cloud platforms (AWS, Azure, GCP) and backend frameworks including Django and FastAPI. Passionate about open source, hackathons, and building scalable, production-ready systems.
Work Experience
- •Sole engineer on the MegaLLM backend — a production OpenAI/Anthropic-compatible API gateway in front of 70+ LLMs on Hono + Bun + TypeScript with MongoDB, Redis, Kafka, and ClickHouse.
- •Designed a 3-tier cache (in-memory LRU → Redis → MongoDB) hitting >95% L0 hit rate at ~2–4ms, plus intelligent model routing with a MobileBERT classifier and 429/5xx fallback chains.
- •Shipped Stripe + Razorpay + OxaPay billing with idempotent webhooks, organization wallets, and ClickHouse credit-event streaming for usage correlation.
- •Hardened the platform with JWT + Redis blacklist, prefix-lookup API keys, 3-tier admin RBAC, and full Prometheus/OpenTelemetry/Grafana Loki observability.
- •Co-founded and engineered an AI-powered meeting platform with real-time transcription and multilingual translation
- •Designed low-latency AI services using Python, FastAPI, and WebSockets for real-time audio streaming
- •Improved transcription stability through streaming optimizations and async processing pipelines
Projects
Govt. Pre Sabka AI Hackathon — 1st Position
Award-winning full-stack personal safety platform with real-time GPS tracking, AI distress detection, geofenced safe zones, covert duress-password flow, and a government dashboard for jurisdiction-based emergency response.
- Won 1st Position at the Govt. Pre Sabka AI Hackathon
- Built the FastAPI + PostgreSQL backend with JWT (httpOnly cookies), bcrypt, SQLAlchemy ORM, SlowAPI rate limiting, and Sentry error tracking
- Implemented a 5-second cancellable SOS countdown with parallel dispatch to SMS, WhatsApp, email, and voice via Twilio, plus token-based public live-tracking pages
- Integrated multi-provider AI — Whisper/Deepgram for transcription, Claude (via MegaLLM) for distress analysis, Azure OpenAI Realtime for the AI Safety Call feature, and ElevenLabs for TTS
- Shipped geofencing with Haversine calculations to auto-start/stop walk sessions, plus a Next.js 15 PWA frontend with Leaflet maps and Zustand state
Smart India Hackathon
Developing a full-stack platform using Node.js, Express, MongoDB, and Next.js to improve college enrollment through personalized recommendations.
- Implemented secure authentication, role-based access control, and RESTful APIs for courses and recommendations
- Integrated interactive college maps and multilingual content delivery using translation APIs
Personal Project
Developed a Django-based platform for conducting quizzes and mock exams with secure login and result tracking.
- Implemented instant scoring, answer review, and an admin panel for managing tests, questions, and users
Public Project
Developed and deployed a multifunctional Discord bot with automation, moderation, and engagement features.
- Built REST APIs and a Flask-based Card Graphics API for dynamic profile cards and leaderboards
- Scaled the system to 635 guilds and 1.3M+ users while maintaining reliable backend performance
Education
B.E. in Computer Science
Thapar Institute of Engineering & Technology
Higher Secondary School (77.2%)
SSD Krishna Vatika School
Secondary School (81.8%)
Bishop Cotton School
Technical Skills
Languages
Databases
Backend
Frontend
Cloud & DevOps
Observability
Core CS
Tools
Problem Solving & Competitive Programming
Regular practice on LeetCode and hackathons, focusing on problem-solving with Data Structures & Algorithms.
Proficient in arrays, strings, hash tables, linked lists, stacks, trees, tries, sorting, dynamic programming, divide & conquer, and bit manipulation. Strong foundation in applying algorithmic techniques to solve real-world problems efficiently.