
Rajath Kumar
Edge AI Engineer & Founder, Analog Data
Bengaluru, India
I build things that run on chips and the software that talks to them. ESP32, STM32, FreeRTOS, FastAPI, TinyML — from bare-metal firmware to cloud backends to on-device inference. Based in Bengaluru. Founder of Analog Data.
Posts by Rajath Kumar(17)
Building AI Agents with LangGraph: State Machines for LLM Workflows
AI agents aren't just prompt chains. LangGraph brings state machines, conditional routing, and human-in-the-loop to LLM workflows for production agents.
LoRA Fine-Tuning for Edge Deployment: Shrink, Quantize, Ship
Full fine-tuning is expensive and wasteful when you only need domain adaptation. LoRA trains 1% of the parameters, quantizes to INT8, and runs on edge hardware.
Deploying LLMs with vLLM and Docker: A Production LLMOps Guide
Serving LLMs in production isn't just loading a model in Flask. Learn vLLM for high-throughput inference, Docker for reproducible deploys, and the ops layer.
Real-Time Edge Vision: OpenCV + ESP32-CAM for Embedded Inference
Stream video from an ESP32-CAM, process frames with OpenCV on the edge, and run lightweight inference — all without a cloud dependency.
Advanced RAG: Hybrid Search, Reranking, and Citation for Production
Basic RAG retrieves and generates. Production RAG uses hybrid search, cross-encoder reranking, and grounded citations. Here's how to build the latter.
PyTorch Training Fundamentals: From Tensors to Custom Datasets
Stop copy-pasting PyTorch boilerplate. Understand tensors, autograd, datasets, and training loops from the ground up — the way an engineer should.
ESP32 Dual-Core Task Pinning with FreeRTOS: A Production Guide
Most ESP32 tutorials ignore dual-core pinning entirely. Here's how production firmware uses both cores — with real task allocation patterns from the field.
MQTT vs HTTP for IoT: Which Protocol Should Your Firmware Use?
Most IoT tutorials default to HTTP without explaining why. Here's a head-to-head comparison of MQTT and HTTP from a firmware engineer who deployed both.
ESP32 Memory Architecture: IRAM, DRAM, Stack, and Heap Explained
Stack overflows are the #1 silent killer of ESP32 production firmware. Understanding IRAM, DRAM, stack and heap allocation is the foundation of stable firmware.
TinyML on ESP32: Building an Anomaly Detection Model from Scratch
Running a trained anomaly detection model on an ESP32 with no cloud. Here's the full pipeline — data collection, training, quantization, and inference.
Connecting ESP32 to AWS IoT Core: A Production Setup Guide
Most AWS IoT tutorials skip the hard parts — certificates, device shadows, reconnect logic. Here's the production setup that survives in the field.
ESP32 and FreeRTOS: Why Your Arduino Loop Is Holding You Back
The Arduino loop() runs one thing at a time. FreeRTOS runs many — simultaneously, reliably, with priorities. Here's why every serious ESP32 project needs it.
Running LLMs Locally: The Engineer's Practical Guide to Ollama
No API keys, no cloud costs, no data leaving your machine. Ollama makes running LLMs locally practical for engineers who want real AI integration.
The Embedded Systems Engineer Career Roadmap for India in 2026
The Indian embedded systems job market is growing fast — driven by EV, defence, IoT, and Edge AI. Here's the skills roadmap, salaries, and companies hiring.
Building a RAG Pipeline for Engineers: From PDF to Answers in 50 Lines
RAG sounds complicated. It's not. Here's how to build a working pipeline that answers questions from your own documents in under 50 lines of Python.
Build Log #01: Why I'm Building Analog Data and What Almost Stopped Me
Six months in, one cohort complete. Here's what I got wrong, what I almost quit over, and what the data showed about embedded systems education in India.
Building a Production FastAPI App with SQLAlchemy
A practical guide to building a FastAPI application with SQLAlchemy models, Pydantic schemas, and route handlers — organized into tabs for easy navigation.