Hire the Best RTL Specialists in Mumbai, IN

I am an Electronics Engineer from IIT Indore with a strong foundation in firmware development and RTOS. My experience includes working with proprietary radio stacks, BLE, and LTE modems for IoT applications. Additionally, I have deployed computer vision Edge AI models on the latest MCUs, significantly reducing the BOM cost while optimising performance across embedded systems.

Full-Stack Hardware & Software Engineer | Embedded Systems | FPGA | PCB | Cloud | AI/ML | Mobile Apps I build complete products from silicon to cloud as a 10-year electronics and embedded systems professional. Hardware & Firmware: PCB Design → FPGA (Verilog/VHDL) → Embedded C/C++ → VLSI → Microcontrollers → Telecom Systems Software & Cloud: AWS Cloud Architecture → Flutter (Android & iOS) → Mobile App Development → Backend Integration → IoT Pipelines AI & Emerging Tech: Machine Learning on Edge Devices → LLM Integration → AI-Powered Embedded Systems → TinyML I deliver end-to-end solutions for small and medium businesses from schematic to shipped product, firmware to mobile app, hardware to cloud dashboard 🏆 Rajya Puraskar Awardee | State Volleyball Champion discipline, focus, and never-give-up attitude drive every project I take on. What you get when you hire me: One engineer who covers the full stack no handoff delays, no miscommunication between hardware and software teams, faster delivery, lower cost. Let's build something remarkable.

I Build Voice-AI Systems That Actually Run in Production - (inc call branding) I design and deploy real-time voice AI and LLM systems that handle live customer conversations, automate phone workflows, and connect directly into enterprise systems — at scale, with low latency, and without black-box dependencies. This is not demos. These are production systems running today. What I Build (Concrete Outcomes) Voice AI Agents That Handle Calls End-to-End Answer inbound phone calls Qualify leads in real time Route, schedule, escalate, or resolve Trigger downstream systems automatically (CRM, billing, ops) Built for sub-second latency, natural turn-taking, and reliable audio. Real-Time Conversational Systems (Voice + Text) Live transcription and streaming responses Contextual memory across turns Session-safe, tenant-isolated conversations Voice and chat ingress/egress using the same intelligence layer LLM-Driven Reasoning with Real Data LangChain-based orchestration RAG pipelines with Pinecone / Weaviate Tool calling into internal APIs Deterministic guardrails and structured outputs LLMs are used as reasoning engines, not toys. Telephony and Streaming Infrastructure You Control Twilio, Telnyx, SIP LiveKit (self-hosted or cloud) WebRTC media pipelines PSTN ↔ WebRTC bridging Full ownership of data, audio, and logs No vendor lock-in. No mystery routing. Workflow Automation That Actually Connects n8n pipelines triggered directly from calls LangChain + MCP integrations Data ingestion, enrichment, and action in real time Voice → tools → systems → voice loop Your phone system becomes an execution engine. Why Clients Hire Me (Reality, Not Promises) Sub-second audio latency under real load Stable voice streaming at hundreds of concurrent calls Production-grade architecture, not notebooks Self-hosted, auditable, secure deployments Clear separation of media, AI, and control planes Systems designed to scale from day one I don’t experiment on production traffic. Live Systems Running Today These are not concepts — they have users right now. OneClickDeploy LiveKitHelp VoiceAIPulse Technical Depth (What This Is Actually Built On) Real-time media: LiveKit, WebRTC, SIP, RTP Telephony: Twilio, Telnyx AI orchestration: LangChain, Pipecat, MCP LLMs: OpenAI, Claude RAG: Pinecone, Weaviate Backends: Python, FastAPI, Node.js Infra: Docker, Kubernetes, Redis Automation: n8n Performance: GPU deployments (CUDA, TensorRT) I understand the entire stack — from audio packets to LLM reasoning to CI/CD. Positioning (Clear and Honest) If you need: A demo → not a fit A chatbot → not a fit A “maybe it scales” system → not a fit If you need: 500+ concurrent calls Deterministic behavior call branding so it doesn't get filtered as "spam" Low-latency voice Full infra ownership Clean scaling and failure isolation That’s exactly what I build.