SOUMABRATA GHOSH
Backend & Full Stack Engineer
PROFESSIONAL PROFILE
Self-taught backend and AI engineer who builds depth-first, from-scratch systems to internalize how real infrastructure works — spanning distributed systems, database internals, and backend architecture. Comfortable working across the stack, with particular strength in system design, concurrency, and data-layer correctness. Currently targeting a Software Engineering role.
CORE SKILLS
Languages: Python, TypeScript, JavaScript, SQL, Java
Backend & Data: Node.js, NestJS, FastAPI, PostgreSQL, Redis, Kafka, Prisma
Frontend: React
Systems & Concepts: Distributed systems, WAL & crash recovery, Raft-like consensus, database internals, concurrency, Docker, CI/CD basics
AI/ML: ML foundations (linear algebra), Gemini API, Chroma, RAG-style application design
SELECTED PROJECTS
WAL-Based Replicated Key-Value Store
Independent Project
- Designed a write-ahead-log-based key-value store with Raft-like replication semantics.
- Worked through core distributed-systems design questions: LSN persistence, AppendAck timing, heartbeat vs. empty AppendEntries, and WAL truncation after snapshotting.
E-Commerce Backend
Independent Project
- Built a production-style e-commerce backend with NestJS, Prisma, PostgreSQL, Redis, BullMQ, and Kafka, scoped from a detailed PRD and ER diagram.
Webhook Relay Service
Independent Project
- Built a Python webhook relay with HMAC auth, rate limiting, and JMESPath-based filtering.
- Implemented a transactional outbox pattern with SELECT FOR UPDATE SKIP LOCKED and a Redis-backed circuit breaker, verified with a chaos-testing suite.
DNS Server & Raw HTTP Server
CodeCrafters / Independent Project
- Implemented a DNS server from scratch in Python (wire format, bit-packing, pointer compression, forwarding proxy), verified end-to-end against a live domain.
- Built a raw HTTP server in Python using only the socket and threading modules.
Payment System Integration
Independent Project
- Built a payment system integrating Cashfree with TypeScript, Node.js, and Prisma, with idempotency handling for safe retries.
LEARNING & PRACTICE
- Reading Designing Data-Intensive Applications (DDIA) as a backbone for project selection and systems concepts.
- Studying ML foundations (linear algebra for ML) and regularly practicing Python and Java DSA / LeetCode problems.
- Deep hands-on study of PostgreSQL (window functions, recursive CTEs, joins, aggregates) and distributed messaging (Kafka, Redis pub/sub, RabbitMQ, sharding strategies).