Somya Prasad
Backend engineer with 4+ years building distributed, event-driven systems. Specialized in LLM-integrated apps, RAG, and real-time pipelines with Kafka, MQTT, and PostgreSQL—production systems at 10K+ concurrent device streams and AI-powered workflows.
From frontend and APIs to distributed backends, RAG, and real-time data at scale
Asentria · Bengaluru
Real-time MQTT ingestion for 10K+ concurrent device streams; Python telemetry pipelines; Kafka + Redis event-driven architecture; PostgreSQL schema and indexing work (~60% faster queries on hot paths); AWS operations. Built a RAG conversational interface over SiteBoss device specs—hybrid retrieval, grounded LLM responses, streaming APIs, and embedding cache.
Morfdesk
Backend for LLM-driven Shopify automation (Node.js, Prisma). RAG with embeddings and vector search; OpenAI integrated with relational data; Kafka pipelines for async LLM jobs; Redis cut inference latency ~45%.
Fuze India · Bengaluru
Distributed backend modules with gRPC and Redis (99.95% uptime). Scaled MySQL for 100K+ ride records/day; Redis caching reduced load ~55%; AWS auto-scaling for production.
Karewise · New Delhi
SEO lifted landing traffic ~45%. React and Material UI; performance with Context, useMemo, and useCallback (~40% fewer re-renders); lazy loading and route-based code splitting (~60% faster initial load). Node.js REST APIs and Redis server-side caching (~30% faster APIs).
Tidbeat · Hyderabad
Migrated a large React codebase to Next.js (SSR/SSG; ~25% Lighthouse SEO gain). GraphQL integration; WhatsApp Business API for support; real-time delivery tracking with Google Maps and Geolocation APIs.
BMSIT · Bengaluru
Undergraduate degree. Also completed Python fundamentals certificates (Coursera): Python Basics and Python Functions, Files and Dictionaries.
A comprehensive toolkit spanning full-stack development to cutting-edge AI technologies
Selected work across RAG, real-time data, and distributed backends
Hybrid retrieval with PostgreSQL filters plus vector similarity; log-to-root-cause pipeline using embeddings and LLM inference; Kafka ingestion for real-time context; natural-language APIs over structured and unstructured telemetry.
Retrieval-augmented conversational interface for SiteBoss capabilities—domain embeddings over product specs, hybrid retrieval with structured filters, grounded LLM answers for sensor configs and telecom deployment guidance, plus streaming query orchestration and cached context.
Backend systems for LLM-driven Shopify workflows: Prisma and Node.js, RAG pipelines, OpenAI with relational data, Kafka-triggered async jobs, and Redis to cut repeated-inference latency ~45%. Shipped multiple production storefronts.
Live sites
gRPC and Redis services with strong concurrency; MySQL scaled for 100K+ ride records per day; caching dropped backend load ~55%; AWS auto-scaling for reliability.
Landing SEO (+45% traffic), responsive React and Material UI, performance tuning with Context and memoization, Node.js auth APIs, and Redis-backed APIs (~30% faster responses).
Ready to build something extraordinary? Let's make it happen
Backend engineer with 4+ years building distributed, event-driven systems. Specialized in LLM-integrated apps, RAG, and real-time pipelines with Kafka, MQTT, and PostgreSQL—production systems at 10K+ concurrent device streams and AI-powered workflows.