Daniel Melnikov

R&D Engineer

IoT Embedded Edge ML Physical AI

I build the software and hardware needed to understand and interact with the real world, enabling systems that tie sensors to decisions.

What I Work On

Embedded

Bare-metal and RTOS on STM32, ESP32, PowerPC. Mixed signal board design. Boot chip-ops to ultra-low power remote nodes.

Edge ML

On-device inference with TFLite Micro and PyTorch. Detection and classification where latency and power budgets are tight.

IoT

Sensing, compute, connectivity (BLE, WiFi, MQTT, LoRa), and the software that makes networked devices useful.

Physical AI

Systems that model the physical world and act on that information: perception and response in a closed loop.

Projects

Eurorack Modular Synthesizer

I design and build synthesizer modules across three paradigms: MCU firmware (DSP on embedded platforms), FPGA gateware (audio in HDL), and analog circuits (hand-soldered on protoboard). It's a personal lab for exploring how analog, digital, and programmable logic interact through voltage as a shared interface. My current project, Movement Voltage, tracks a hand with a stereo camera and turns gesture into control voltage: the body becomes the instrument.

Demoscene Sizecoding

I have made visual demos in ≤140 characters of JavaScript. Starfields, matrix rain, abstract geometries, procedural textures: all generated from extremely compressed math and syntax. The constraint forces you to find new solutions to problems you thought you understood and recombine what you've learned to squeeze more out of fewer characters.

About

I enjoy giving “life” to continuously functioning things. That could be signal training for a 25.6GT/s memory controller, bringup for a mainframe processor built to eight nines of uptime, code golfing a chaotic data sculpture, or development of a synthesizer module that blurs the line between human and machine.

Contact