You will get Custom TinyML & Edge AI Solutions for Microcontrollers
Rising Talent
Rising Talent
Project details
Bring Intelligence to the Edge: Custom TinyML & Embedded AI Solutions.
Stop sending raw data to the cloud. I engineer optimized AI models that run directly on microcontrollers, delivering real-time decisions with zero latency and 100% data privacy. As the founder of EdgeForge Labs, I specialize in squeezing high-performance Machine Learning into low-power hardware like ESP32 and STM32.
Why choose this service?
Privacy First: Process sensitive data (voice, video, biometrics) locally on the device. No data leaves the chip.
Ultra-Low Latency: Millisecond reaction times for critical industrial or robotics applications.
Full-Stack Implementation: I don't just train the model in Python; I deploy it in efficient C++ firmware optimized for your specific MCU.
Advanced Optimization: Expert in Model Quantization (int8) and Pruning to fit complex neural networks into limited Flash/RAM without sacrificing accuracy.
Whether you need predictive maintenance, keyword spotting, or object detection, I turn "dumb" devices into intelligent assets.
Stop sending raw data to the cloud. I engineer optimized AI models that run directly on microcontrollers, delivering real-time decisions with zero latency and 100% data privacy. As the founder of EdgeForge Labs, I specialize in squeezing high-performance Machine Learning into low-power hardware like ESP32 and STM32.
Why choose this service?
Privacy First: Process sensitive data (voice, video, biometrics) locally on the device. No data leaves the chip.
Ultra-Low Latency: Millisecond reaction times for critical industrial or robotics applications.
Full-Stack Implementation: I don't just train the model in Python; I deploy it in efficient C++ firmware optimized for your specific MCU.
Advanced Optimization: Expert in Model Quantization (int8) and Pruning to fit complex neural networks into limited Flash/RAM without sacrificing accuracy.
Whether you need predictive maintenance, keyword spotting, or object detection, I turn "dumb" devices into intelligent assets.
AI Development Type
Deep Learning, Model TuningAI Tools
MATLAB, NVIDIA AI Platform, OpenCV, PyTorch, TensorFlowAI Development Language
PythonWhat's included
| Service Tiers |
Starter
$500
|
Standard
$1,500
|
Advanced
$2,500
|
|---|---|---|---|
| Delivery Time | 7 days | 21 days | 45 days |
Number of Revisions | 2 | 3 | 5 |
AI Model Integration | - | - | |
Detailed Code Comments | - | ||
Knowledge Graph | - | - | - |
Model Documentation | |||
Ontology | - | - | - |
Source Code | - | ||
Taxonomy | - | - | - |
Optional add-ons
You can add these on the next page.
Fast Delivery
+$100 - $500
Additional Revision
+$50
Dataset Augmentation
(+ 3 Days)
+$250
Live Flashing Support
+$100Frequently asked questions
4 reviews
(4)
(0)
(0)
(0)
(0)
This project doesn't have any reviews.
CB
Christopher B.
Apr 24, 2026
Engineering Support: Bill of Materials (BOM) Validation and Multi-Factory Production Setup
Another successful project with the help of Jhordan. Jhordan produces quality work in a timely fashion while clearly communicating any hurdles as they are encountered. This project covered gathering quotes from 3 fab houses, and producing 5 prototype boards for a USB-PD module. From start to finish, it is a breeze working with Jhordan and I continue to look forward to working on further projects together.
CB
Christopher B.
Apr 24, 2026
KiCad PCB Revision + JLCPCB Production Preparation (Existing Board)
Working with Jhordan continues to be a pleasure. Jhordan produces quality work in a timely fashion while clearly communicating any hurdles as they are encountered. This project covered some chip replacements on an existing design, gathering quotes from 3 fab houses, and producing 5 prototype boards. From start to finish, it is a breeze working with Jhordan and I continue to look forward to working on further projects together.
CB
Christopher B.
Mar 28, 2026
RP2040 Firmware Serial Communication Framing
Jhordan has helped me with number of project scopes so far. He is detail oriented, speedy, and checks in regularly. I plan to work with him regularly here onward.
CB
Christopher B.
Jan 16, 2026
KiCad USB-PD Modular Sub-Board Design
Jhordan did a fantastic job at separating out a module from our schematic design and PCB layout. Jhordan communicated very well, and confirmed several details before proceeding with any significant amount of work. Updates were given frequently and the work finished well within the expected timeframe. I'll be working with Jhordan more in the very near future.
About Jhordan
Full-Stack Hardware Engineer | IoT - TinyML - IIoT SCADA
Quito, Ecuador - 12:56 am local time
At EdgeForge Labs, I close that gap.
I'm Jhordan, an Electronics & Automation Engineer specializing in end-to-end intelligent hardware systems. My work spans the full stack: schematic capture and PCB layout in KiCad (DFM-optimized for JLCPCB/PCBWay production), embedded firmware in C/C++ on STM32 (HAL/LL/FreeRTOS) and ESP32 (ESP-IDF), IoT connectivity (BLE, MQTT, LoRaWAN, OPC UA), and — where performance demands it — TinyML models deployed directly on the microcontroller.
What I build:
→ IoT Product Prototypes: From concept to a manufacturable, firmware-complete hardware prototype. I handle component selection, schematic, PCB layout, BOM management, and bring-up firmware — all designed for the day your product scales beyond the prototype.
→ TinyML & Edge AI Systems: I deploy optimized ML inference directly on STM32 and ESP32 — no cloud dependency, millisecond latency. Projects include predictive maintenance (RUL estimation from vibration data), neural network control systems, and real-time anomaly detection. I handle the full pipeline: dataset preparation, model training in Python, quantization (int8), and deployment in efficient C++ firmware.
→ Industrial IIoT & SCADA Modernization: I connect legacy PLC infrastructure (Siemens, Rockwell, Schneider) to modern IT systems via OPC UA, Modbus TCP, and PROFINET. I design high-availability SCADA architectures, implement redundant PLC control schemes, and build real-time monitoring dashboards that follow ASM (Abnormal Situation Management) standards.
My engineering standard:
Every deliverable is production-oriented. That means copper pullback specs that don't get Gerbers rejected, firmware architectures that survive field resets, and SCADA systems designed with redundancy before the client asks for it.
I work in English and Spanish. My communication style is direct: I flag risks early, document clearly, and propose solutions before you have to ask.
If your project sits at the intersection of hardware, firmware, and intelligence — let's talk.
Steps for completing your project
After purchasing the project, send requirements so Jhordan can start the project.
Delivery time starts when Jhordan receives requirements from you.
Jhordan works on your project following the steps below.
Revisions may occur after the delivery date.
Data Strategy & Architecture
I analyze your dataset quality (or design a collection plan) and select the optimal Neural Network architecture compatible with your hardware constraints.
Model Training & Optimization
I train the model (TensorFlow/Edge Impulse) and perform Quantization/Pruning to compress it, ensuring it fits within the limited Flash/RAM of the MCU.