AI Tools Used in Engineering

When I think of embedded systems, robotics, industrial automation, marine, energy, IoT, etc.,  I’m curious what AI tools engineers are using and trying out.  In talking with engineers from various marine tech companies, AI experts, and the web, I see some meaningful impact on real engineering work.

1. Cadence Cerebrus (AI-Driven EDA)
Category: Right down to the bare metal, System-On-Chip design and dense PCBs.
Why it matters:
Cerebrus is one of the first large-scale AI engines baked into an EDA suite. It uses reinforcement learning to optimize:
· layout
· power
· thermal profiles
· timing
· signal integrity
Engineers get lower risk, faster iteration, fewer layout errors, and dramatically improved design closure.
Best for: Teams doing high-performance embedded designs, custom silicon, or dense PCBs.

2. MathWorks MATLAB/Simulink with AI Toolboxes
Category: Embedded systems, control systems, signal processing
Why it matters:
MATLAB remains the workhorse of signal processing and control engineering. Add AI toolboxes and you get:
· AI-enhanced signal filtering
· Predictive maintenance modeling
· Neural-network-based control strategies
· Faster model-based design validation
· Auto-generation of embedded C/C++ for microcontrollers
Best for: Any engineer combining sensors + firmware + control loops + signal analysis.

3. Edge Impulse
Category: Edge AI / TinyML for embedded hardware
Why it matters:
This is the go-to platform for getting machine learning running on microcontrollers and low-power devices.
Engineers use it for:
· anomaly detection
· vibration monitoring
· audio/voice triggers
· powerline anomaly detection
· onboard sensor classification
· predictive maintenance
It integrates cleanly with STM32, Arduino Pro, Nordic, TI, ESP32, etc.
Best for: Hardware that needs lightweight ML models running directly on the device — without cloud dependencies.

4. GitHub Copilot / AI Code Assistants (with hardware-aware extensions)
Category: Firmware, embedded software, diagnostics
Why it matters:
Copilot (and similar tools) have quietly become indispensable productivity tools for teams writing:
· embedded C/C++
· FreeRTOS tasks
· control loops
· I/O drivers
· CAN, Modbus, SPI, I2C layers
· diagnostic routines
· test harnesses
· simulation scaffolding
They don’t replace firmware engineers — they speed up development and reduce bug counts.
Best for: Any environment where firmware and hardware interact tightly.

Other tools worth trying:

• NVIDIA Jetson + TAO Toolkit

For robotics, machine vision, and edge inference.

• Keysight PathWave (AI-powered test analytics)

Super strong for lab + production hardware validation.

• Ansys AI plugins

AI-assisted thermal, EMI/EMC, vibration, and multiphysics.

• IBM Maximo / Falkonry AI

Used for industrial electrical predictive maintenance.