Sureshkumar R

Full Stack Engineer · AI Systems & Multi-LLM Agent Developer · PWA Specialist

Full Stack Engineer with 3+ years of experience building scalable, high-performance web applications, multi-LLM agent workflows, and offline-first PWAs. Currently developing OneReach (OmniReach AI) at OneData Software Solutions, focusing on multi-channel sales outreach, agentic AI research pipelines, and cloud infrastructure. 🚀 Proven Track Record: Delivered enterprise & government software adopted across 2 state governments, serving 10,000+ active students. 🤖 AI Systems Engineering: Experience orchestrating stateful multi-agent LLM pipelines (LangGraph, Claude 3.5 Sonnet, GPT-4o, Gemini, AWS Bedrock), vector search (pgvector), and voice AI WebSockets. ⚙️ Performance & Architecture: Reduced JS bundle sizes by 82% (2.5 MB → 450 KB), optimized rendering speed by 60%, and saved $4,200+ in production API spend.

• ⚡ Performance & Scale First: 82% bundle size reduction (2.5 MB → 450 KB), sub-second TTFB, and offline-first service worker caching for 10,000+ users across low-connectivity regions.
• 🛡️ API Cost & Resilience: Designing fail-safe caching layers, rate-limiting, and stateful fallback execution—preventing $4,200+ in production API waste.
• 🎯 Component Architecture & Standards: Atomic Design Systems (50+ reusable React/TS components) engineered strictly around WCAG 2.1 AA accessibility guidelines.

💼 Full Stack Engineer @ OneData Software Solutions                [Jun 2026 – Present]
   └── Core developer on OneReach (OmniReach AI) — Multi-channel sales outreach & AI research platform

💼 Software Engineer (SDE-II) @ NavaDhiti Business Consultancy   [Jul 2023 – Mar 2026]
   └── Scaled 3 concurrent platforms (AXL, Wildway, TELL) for client EkStep Foundation

Jun 2026 – Present | Coimbatore, India

• Multi-Tenant Platform & Backend: Architected OneReach (OmniReach AI) using Node.js (ESM), Express, Drizzle ORM, and PostgreSQL (pgvector) with pg-boss queues, automating LinkedIn, Email, SMS, and WhatsApp messaging.
• Multi-LLM Agent Workflows: Orchestrated autonomous prospect discovery & enrichment using LangGraph, Claude 3.5 Sonnet, GPT-4o, Gemini, and Apollo API.
• Voice AI & Cost Engineering: Built real-time WebSocket voice AI bridge, React Native mobile apps, and Chrome extension; saved $4,200+ in production API spend via intelligent caching.

Jul 2023 – Mar 2026 | Bangalore, India

• State-Scale Platforms: Delivered 3 production PWAs (AXL, Wildway, TELL) for EkStep Foundation, adopted by 2 state governments (Karnataka & Telangana) for 10,000+ active students.
• Design System & Performance: Architected a 50+ component React/TypeScript Atomic Design System and engineered an 82% JS bundle reduction (2.5 MB → 450 KB) via code splitting and CDN asset offloading.
• Offline-First PWA: Achieved 60% faster rendering on low-connectivity networks using Service Workers and progressive caching strategies.
• Microservices Architecture: Designed 6 independently deployable GCP microservices (orchestration, content, learner, telemetry) maintaining ISO 27001 compliance.

💻 Frontend & Mobile	⚙️ Backend & Data Architecture
🤖 AI Systems & LLM Orchestration	☁️ DevOps, Cloud & Testing

Multi-channel messaging engine, LangGraph agent research pipeline & Voice AI WebSocket bridge

┌─────────────────────────────────────────────────────────────────────────────────────────┐
│ React 18 / Vite / TS  ──► Node.js ESM Backend (Drizzle + pgvector) ──► Multi-Tenant DB   │
│ React Native / Chrome ──► WebSocket Voice AI Bridge ──► LangGraph Multi-LLM Agents      │
└─────────────────────────────────────────────────────────────────────────────────────────┘

• Multi-Channel Automation: Unified LinkedIn, Email (Gmail/Outlook), SMS & WhatsApp messaging engine with strict multi-tenant workspace isolation (pg-boss queues).
• Agentic AI & Enrichment: Stateful LangGraph multi-agent pipelines for lead discovery & Apollo API company data enrichment.
• Real-Time & Cross-Platform: Low-latency WebSocket Voice AI bridge, Chrome extension for LinkedIn research, and React Native offline lead capture.
• Cost Engineering & Impact: Saved $4,200+ in production API spend via intelligent query caching & rate limiting.

React 18 TypeScript Node.js (ESM) Express Drizzle ORM PostgreSQL pgvector LangGraph OpenAI Claude 3.5 AWS Bedrock WebSockets React Native Chrome Extension TailwindCSS Docker Stripe Playwright

Offline-first educational PWA & analytics dashboard · EkStep Foundation

• Built state-level learning PWA and administrative analytics dashboard deployed across Karnataka & Telangana serving 10,000+ active students.
• Deployed across 6 GCP microservices (orchestration, content, telemetry) with 80%+ unit test coverage using Jest/RTL.
• Achieved an 82% JS bundle reduction (2.5 MB → 450 KB) and reduced administrative manual reporting time by 45%.

React.js Next.js Node.js MongoDB PWA GCP Microservices Recharts

Geospatial field-tracking PWA & automated alert trigger pipeline · EkStep Foundation

• Developed real-time incident logging platform for wildlife field rangers operating in remote, low-connectivity zones.
• Integrated Firebase Realtime DB sync and push notifications, boosting ranger incident response efficiency by 50%.

React Firebase Realtime DB PWA Push Notifications Gemini API

Multilingual speech recognition & pronunciation feedback platform · EkStep Foundation

• Integrated Bharat ASR/TTS APIs across 10+ Indic languages for spoken pronunciation assessment serving 5,000+ learners.
• Leveraged Indic-language phoneme alignment, outperforming generic ASR systems in spoken reliability and scoring precision.

React Node.js Bharat ASR/TTS Indic Phonemes Adaptive Analytics

• 🤖 DeepSeek R1 & Local LLMs: Benchmarking quantized open-weight reasoning models via Ollama & vLLM for zero-cost local agent fallback.
• 🧠 Advanced RAG & Hybrid Search: Implementing HNSW indexing & reciprocal rank fusion (RRF) combining BM25 full-text search with pgvector embeddings.
• ⚡ Edge AI & WebGPU: Experimenting with browser-side AI models (Transformers.js) for instant offline speech & text evaluation.