Headline

Founding Software Engineer | Backend & AI Systems | Node.js, Python, Apache Kafka, Microservices, AWS | GenAI, RAG, pgvector | @ Dotsin (Intel AI Partner) | Ex-Assessli | Bengaluru & Remote

About

Backend Distributed AI Systems Engineer with proven experience architecting scalable monorepo microservices, event-driven streaming pipelines, and production GenAI infrastructure.Currently building at Dotsin (Intel AI Partner), where I co-architected a 15-microservice platform, stabilized high-throughput Apache Kafka event pipelines, and reduced LLM inferencing costs by up to 40x using OpenVINO vLLM optimizations. Previously built fault-tolerant evaluation pipelines at Assessli ($5M funded EdTech).CORE COMPETENCIES:• Backend Architecture: Node.js (TypeScript), Python (FastAPI), 15+ Microservices, REST gRPC• Event Streaming Data: Apache Kafka (Aiven, SCRAM/SSL), Redis, PostgreSQL, pgvector• AI ML Systems: LLM Orchestration, RAG Pipelines, vLLM, Intel OpenVINO, Thompson Sampling Bandits• Cloud DevOps: AWS (EC2, S3, IAM), GCP (Cloud Tasks, Pub/Sub), Docker, Turborepo MonorepoLooking to drive high-impact architecture as a Backend, AI Systems, or Founding Engineer across high-growth startups and tech teams.📍 Bengaluru Remote

Location

Kolkata

AI stack

antigravity

claude code

claude

codex

Work experience

Software Engineer at Dotsin, May 2026 - Present · 6 mos, Bengaluru, Co-architected a pnpm + Turborepo monorepo with 15 independently deployable services (14 TypeScript/Express, 1 Python/FastAPI). Stabilized Kafka-based event infrastructure (Aiven, SASL SCRAM-SHA-256 over SSL) powering 4 critical consumer groups for discovery and engagement loops. Resolved a production consumer-eviction issue and reduced projected LLM/API waste by 30-50x via timeout tuning and in-memory deduplication. Built hybrid feed ranking with dense retrieval, Thompson Sampling contextual bandits, and Gemini reranking with live reward updates.

Software Developer at Assessli, Aug 2024 - Apr 2026 · 1 yr 9 mos, Kolkata, Architected a market-first AI grading engine integrating OCR + LLM evaluation for handwritten assessments at human-comparable quality.Built fault-tolerant async processing with GCP Cloud Tasks, Pub/Sub, and Redis queues to absorb high-volume PDF and AI workloads with zero loss.Engineered Gemini-based question generation workflows to automate structured exam creation from document references.

Project Developer Intern at NIT Durgapur, May 2024 - Jul 2024 · 3 mos, Remote

Software Engineer Intern at NIT Durgapur, May 2024 - Jul 2024 · 3 mos, Durgapur, Developed secure Node.js backend services for a rural telemedicine platform delivered through NGO healthcare partnerships. Modeled medicine inventory flows with expiry-aware tracking and supply optimization logic for operational continuity.

Education

Bachelor of Technology, Computer Engineering, Maulana Abul Kalam Azad University of Technology

Bachelor of Technology - BTech, Computer Science Engineering, Budge Budge institute of technology