Ver categorias
Explorar
Fiverr Pro
Português
$
USD
I will design custom tinyml and edge ai applications
Hardware and IoT Specialist
I design PCBs and develop firmware for market-ready hardware. With 160+ successful projects, I specialize in FCC/CE compliant designs that bridge the gap from prototype to mass production.
TECHNICAL ...
Nível 2
Atendeu a critérios de alto desempenho e tem um histórico comprovado de atendimento às expectativas dos clientes.
Sobre este Serviço
Design Custom TinyML & Edge AI Applications (Hardware & Firmware)
Deploy production-ready AI at the edge with custom-built hardware and optimized firmware. I specialize in bridging the gap between complex Neural Processing Units (NPUs) and high-speed PCB design.
Core Specializations:
- AI Hardware: Custom integration of ESP32-S3, STM32 NPU, Coral TPU, Jetson, and Kneron.
- High-Speed PCB: Expert layout for DDR, eMMC, and MIPI-CSI/DSI with strict impedance control and EMI reduction.
- TinyML Firmware: Model quantization and deployment using TensorFlow Lite Micro, ONNX, and STM32Cube.AI.
- Optimization: Advanced thermal management and power delivery networks (PDN) for intensive AI inferencing.
What You Receive:
- Manufacturing Files: Gerber, NC Drill, Pick-and-Place (Ready for JLCPCB/PCBWay).
- Design Assets: Complete Schematics & 3D PCB Models.
- BOM: Full Sourcing list with verified part numbers (DigiKey/Mouser).
- Source Code: Optimized firmware for your specific NPU/accelerator.
Platforms: Jetson , Nicla , Google Coral , ESP32-S3, Hailo AI , NanoEdge AI, OpenVINO, SenseCAP
Note: Every AI proj ect is unique. Please contact me before ordering to discuss your model and hardware requirements.
Clientes com quem já trabalhei
Oasis Water Technologies
Collaborated with multiple teams at Oasis to engineer a highly durable, battery-powered industrial IoT device. Because the device was designed for remote deployments with minimal human interaction, mission-critical reliability was the top priority. I delivered an end-to-end solution, handling the custom hardware design, robust firmware development (including WiFi, OTA updates, and MQTT protocols),
abr. de 2024
Mobile Outfitters
Partnered with Mobile Outfitters to engineer the core electronics for a high-precision, CNC-style mobile accessory cutting machine. Tasked with solving a complex integration of digital and analog circuits, I managed the full hardware lifecycle—from initial schematic design and high-density PCB layout to troubleshooting and prototype testing. The project resulted in a fully manufacturable, producti
fev. de 2025-fev. de 2026
Meu portfólio
Outros serviços de Engenharia Eletrônica que eu ofereço
Perguntas frequentes
Can you optimize my existing PCB for AI inferencing?
Yes. AI chips often require specialized power delivery (PMIC) and thermal dissipation. I can audit your current schematic to include NPUs or TPUs and redesign the layout for better thermal performance and signal integrity.
Do you support TinyML frameworks like TensorFlow Lite?
Absolutely. I design hardware specifically to support TensorFlow Lite for Microcontrollers, Edge Impulse, and ONNX. I ensure the selected MCU has the required SRAM and Flash to handle your specific model size.
What file formats do you provide for manufacturing?
I provide industry-standard Gerber RS-274X files, Excellon drill files, and a comprehensive BOM (Bill of Materials). I also provide 3D STEP files so you can design your enclosure around the AI hardware perfectly.
Can you handle high-speed signals like MIPI-CSI for AI vision?
Yes. Vision-based AI requires careful routing for MIPI-CSI (Camera) and LVDS/DSI (Display). I implement differential pair routing and length matching to ensure your video data remains artifact-free during inferencing.
Is the hardware design ready for FCC/CE certification?
I follow EMI/EMC best practices, including proper grounding planes and decoupling, to ensure your AI product is ready for pre-compliance testing and rapid market entry.
Which AI accelerators do you support?
I design for a wide range of AI-native silicon, including ESP32-S3 (Vector instructions), STM32 with integrated NPUs, and external accelerators like Google Coral (TPU). I match the hardware power to your model's TOPS (Tera Operations Per Second) requirements.
Can you help optimize my AI model to fit the hardware?
While my primary focus is hardware, I design with TinyML constraints in mind. I ensure the PCB has sufficient high-speed PSRAM and Flash to support quantized models (INT8/FP16) from frameworks like TensorFlow Lite or Edge Impulse.
How do you handle high-speed camera or sensor data?
For AI vision or vibration analysis, I implement MIPI-CSI or high-speed SPI/I2S routing with strict impedance control. This ensures low-latency data flow from the sensor to the processor, which is critical for real-time AI inferencing.
Is the design optimized for low-power battery operation?
Yes. Edge AI often lives in remote areas. I specialize in Power Management Integrated Circuit (PMIC) design and "Deep Sleep" hardware triggers to maximize battery life while maintaining the ability to wake up for AI events.
Do you provide the firmware to run the AI model?
I provide the Hardware Abstraction Layer (HAL) and initial "bring-up" code. For full AI model deployment (C++ or MicroPython), please contact me so we can discuss the specific library requirements for your NPU or MCU.
