Backend Engineer · Distributed Systems & AI/LLM Integration · Golang | Kafka | Redis | AWS

Bengaluru, Karnataka, India

Bengaluru, Karnataka, India

Bengaluru, Karnataka, India

Bengaluru, Karnataka, India







Bachelor of Technology - BTech, Computer Science and Engineering
Government College of Engineering, Amravati.
12th, Computer Science
R.K. Jr College
Join Sumit and 10k+ other professionals on
Tal is the app where bosses meet engineers & builders directly, without middlemen
Download Tal BossBackend Engineer · Distributed Systems & AI/LLM Integration · Golang | Kafka | Redis | AWS
Backend engineer with 3+ years building distributed systems, real-time pipelines, and LLM-powered products at scale. I specialize in high-throughput backend systems and AI/LLM integration — from re-architecting SSE pipelines and Kafka-based ingestion systems to shipping conversational bots into production doubt-resolution flows. Some things I've shipped: — ⚡ ~94% CPU reduction by re-architecting an SSE message handling pipeline (45% → 2.5%) 💰 ~$6.4K/month saved by migrating 7 ML models from SageMaker → CastAI on GKE 🔍 Real-time Kafka + Elasticsearch search pipeline powering course discovery at scale 🤖 ~58% bot response coverage via LLM-powered conversational flows in production 🧹 ~29% codebase reduction (19K → 13K LOC) through systematic deprecation Stack: Golang · Java · Python · Kafka · Redis · MongoDB · PostgreSQL · Elasticsearch · AWS · GKE · CastAI · Firebase 📬 tiwaresumit143@gmail.com | sumittiware.netlify.app
Bengaluru
claude code
claude
gemini
langchain
langgraph
mcp
Software Engineer at ALLEN Digital, Jul 2024 - Present · 2 yrs, Bengaluru, Karnataka, India, Doubts Service & LLM Integration: Built a Go microservice for real-time doubt resolution using a hybrid conversational bot — combining intent classification, dialog flow routing, and LLM-first generation with multi-turn session memory. Routed unresolved doubts to human teachers as fallback via an internal LLM gateway, achieving ~58% automated response coverage. Workflow Orchestration Engine: Contributed to an internal config-driven DAG engine (Go/Kratos) — JSON-defined workflows, decision steps, retries, timeouts, and shared context. Designed doubt-resolution pipelines as declarative DAGs orchestrating LLM calls, SSE delivery, persistence, and metrics behind a single API. SSE Pipeline Re-architecture: Identified and eliminated a core bottleneck in the SSE message handling layer — reducing CPU by ~94% (45% → 2.5%), ingress by ~54%, and egress by ~33%, enabling horizontal-scale-free growth. Production Reliability: Built a Go-routine manager with weighted semaphore control and context-based cancellation — eliminating panic vulnerabilities and improving uptime on a high-traffic service. Search Indexing Pipeline: Built and extended an event-driven indexing path — SQS consumer with batching and rate limits, MySQL as staging source-of-truth, and a cron worker bulk-upserting/deleting in OpenSearch. Instrumented end-to-end with OpenTelemetry tracing and metrics. ML Infra Cost Optimisation: Migrated 7 ML models from SageMaker → CastAI on GKE with zero downtime, saving ~$6.4K/month (~55% reduction). Unified QC Platform & Codebase Health: Built a QC platform from scratch replacing fragmented workflows across engineering, product, DevOps, and QA. Separately reduced the backend codebase by ~29% (19K → 13K LOC) through systematic deprecation.
Software Developer at Alippo, Sep 2023 - Jun 2024 · 10 mos, Bengaluru, Karnataka, India, Alippo Search (Elasticsearch) : Designed and built full-text course search from scratch using Elasticsearch, integrated with a Node.js ingestion server that synced data from PostgreSQL in real-time. Powered discovery across a large course catalog. Server-Driven UI (SDUI) Architecture : Designed a backend-controlled UI rendering system enabling over-the-air layout changes without app releases — significantly reducing frontend deployment overhead. In-App Chat Community : Built a WhatsApp-style real-time chat platform supporting 1,000+ DAU across 100+ communities using Firebase real-time messaging. Observability and Monitoring: Set up a Logstash pipeline for centralized log management, enabling proactive monitoring of server errors and application health. Maintained app stability at ~99.6% by proactively monitoring Crashlytics and resolving production issues.
Software Developer at Play GPL, Dec 2022 - Aug 2023 · 9 mos, Bengaluru, Karnataka, India, Fullstack contributor (Golang · Flutter · Next.js) across mobile, backend, and frontend. — Built a Redis-based Feed system and Expert Teams feature (Dream11-style) using an event-driven architecture — Developed an in-house template-based notification system using Go's Firebase Admin SDK with scheduled and automated delivery via go-quartz — Built topic-based user cohort segmentation for targeted notifications — Led mobile app development using MVVM + MobX + go-router
Software Developer Intern at Play GPL, Feb 2022 - Jul 2022 · 6 mos, Bengaluru, Karnataka, India, Startted working at Hoggy later moved to PlayGPL
Bachelor of Technology - BTech, Computer Science and Engineering, Government College of Engineering, Amravati.
12th, Computer Science, R.K. Jr College