Backend engineer graduating from IIT Tirupati (May 2026). I build in production TypeScript codebases: cryptographic protocols, type-safe APIs, query optimization, and security at the system boundary. Currently expanding into AI engineering, exploring LLM tooling and agent infrastructure.

Email: chetanreddyk394@gmail.com

• Production backend systems in TypeScript/Node.js: relay infrastructure, event-sourced architectures, WebSocket servers
• Cryptographic protocol implementation: E2E encryption schemes, key derivation, MAC verification from spec
• Database performance: query batching, schema design, ORM optimization across PostgreSQL and MongoDB
• Type safety and correctness: discriminated unions, Zod validation, co-located auth checks that can't be bypassed
• AI engineering (learning): LLM APIs, agent workflows, prompt engineering, tool use

Languages

Backend

Frontend

Databases

DevOps & Tools

Next.js TypeScript PostgreSQL Prisma Supabase Auth Tailwind CSS v4

• Developers post projects, recruit contributors (capped at 3/project), and apply as mentors
• Role-enforced PostgreSQL schema (owner/contributor/mentor RBAC) with unique membership constraints; repo links gated to approved members only, enforced at the Server Action layer to avoid round-trip latency
• 4-stage GitHub Actions CI pipeline (install, lint/format, type-check, build) with Husky pre-commit hooks scoped to staged files only

TypeScript Node.js WebSockets React shadcn/ui

• Fan-out broadcast iterates only over room members, keeping per-message cost O(room size) vs the O(all connections) naive pattern
• Deliberate design: connection isolation and efficient message routing built in from the start, not bolted on

High-performance relay infrastructure for the Nostr decentralized social protocol

• NIP-44 v2 encryption — Implemented ChaCha20-Poly1305 + HKDF key derivation + MAC verification from scratch, replacing the legacy NIP-04 path that had metadata leaks and weak key derivation
• Relay-boundary access control (NIP-17) — Blocked inner event kinds 13/14/15 at the relay layer regardless of client behavior, preventing unencrypted DMs from ever reaching the DB
• Type-safe CLI refactor — Rewrote runCommandWithOutput as a CommandResult discriminated union, eliminating per-call-site try/catch and making spawn failures type-safe across all CLI commands
• Marmot Protocol support (MIPs 00-03) — Added E2EE group messaging with 5 event kind handlers and a GroupEventStrategy that validates required group-ID tags; relay operates transport-only with no MLS logic
• Performance + validation — Optimized NIP-25 tag validation from multiple .find() calls to a single O(n) pass; migrated all 8 schemas from Joi to Zod, removing the Joi dependency entirely
• Reliability — Fixed a 301 permanent-redirect bug on the relay home page, added XSS prevention via HTML escaping, and in-memory template caching. Raised unit test coverage from ~21% to 88-93% (1,238 passing tests)

Open-source backend powering Bitcoin banking applications

• Event-sourced Wallet module — Replaced mutable DB state with an append-only event log, enabling full audit history and zero-downtime state replay without API contract changes
• N+1 query elimination — Introduced DataLoader batching on the transaction-history endpoint, cutting DB calls from ~60 to ~5 per request (12x reduction) with no changes to the API surface

Exploring the infrastructure side of AI systems:

• LLM API integration (tool use, structured outputs, multi-turn context management)
• Agent architectures and workflow orchestration
• Retrieval-augmented generation (RAG) and vector search
• Prompt engineering and evaluation

Goal: bring the same correctness-first mindset from backend systems into AI application development.

• 🥇 LeetCode Knight — Peak rating 1934, top 3.7%, 500+ problems solved
• ⚡ Juspay Hiring Challenge — Cleared 8 of 9 rounds from a national pool of 200,000+ candidates
• 🏅 Odoo Hackathon — Ranked 41st out of 5,000+ teams
• 🎓 Summer of Bitcoin — Selected from 7,000+ applicants (3-stage competitive process)
• 📊 Adobe Hackathon — Top 4,000 out of 110,000+ teams
• ⭐ CodeChef 3-Star — Peak rating 1610

Targeting junior SDE roles in backend or systems engineering. Open to AI engineering opportunities too. If you're building something where performance, security, and correctness matter, let's talk.