Junior AI Engineer

ZYGO; Zonnebeke; Sur place; EUR 60 000 - 90 000; Plein temps

Avantages offerts par ce poste

Shape ZYGO's future

Latest AI tech

Learning opportunities

Professional workplace

Real-world impact

No traffic

Résumé du poste

ZYGO in Belgium is seeking an AI Engineer with a strong background in multimodal AI across speech, vision, and language models. You will build and deploy on-edge perception and reasoning models on devices like Jetson to enhance ZYGO's perception and interaction with environments.

The role is outcome-driven and offers a path to become AI lead at ZYGO. You will design, train, and optimize on-device models, fuse camera and microphone signals, and collaborate with robotics and cloud engineers to

Qualifications

• Proven experience with deep learning for perception or multimodal reasoning.
• Hands-on experience deploying models on edge devices (Jetson) and optimizing for GPU inference.
• Proficient in Python and deep learning frameworks ( PyTorch, ONNX, TensorRT, CUDA ).
• Ability to build data pipelines, benchmarks and automated tests.
• Thrives in a fast-moving environment and delivers production-ready systems.

Responsabilités

• Design, train and optimize perception and reasoning models that run on-device with low latency and power use.
• Convert and quantize models for Jetson and build surrounding runtime, evaluation and monitoring.
• Fuse signals from cameras and microphones to improve scene understanding.
• Create pipelines, evaluation suites and simulations translating to real-world performance.
• Collaborate with robotics and cloud engineers to ship features into production.

Connaissances

multimodal AI

edge deployment

Python

PyTorch

TensorRT

Outils

PyTorch

ONNX

TensorRT

CUDA

Description du poste

About ZYGO

ZYGO is a fast-growing robotics software company building ZYGOS, a skill-based platform that enables robots to adapt, learn, and operate in complex environments. Our robots are deployed in industries where reliability, autonomy, and intelligence truly matter.

• Most people shouldn’t join a startup. Things break, plans change, and there’s no time for endless meetings or perfect processes. It’s action, every day. If that excites you, you’ll fit right in.

Location

Zonnebeke

Policy

Mostly on-site

Contract

Full-time

Why ZYGO?

• We're still small enough that you'll shape what ZYGO becomes.
• A team that's actively implementing the latest tech to create Physical AI Agents.
• Lots of learning potential with lots of technologies and senior team members.
• Professional workplace despite the fast‑paced growth trajectory.
• Have a real‑world impact on jobs that are under a lot of pressure.
• No traffic jams on your way to work and no wannapreneurs next door.

The role

We’re looking for an AI Engineer with a proven background in multimodal AI across speech, vision, and language models, and a strong hands‑on mindset. You will build and deploy reasoning models on edge devices such as Jetson to advance ZYGOS perception, enabling robots to better understand and interact with their environment. The role is outcome‑driven and offers the opportunity to grow into AI lead at ZYGO.

What You Will Do

• You will design, train, and optimize perception and reasoning models that run reliably on-device, with a focus on latency, power, and robustness in the field.
• You will take models from research to deployment by converting and quantizing them for Jetson using frameworks such as PyTorch, ONNX, and TensorRT, and by building the surrounding runtime, evaluation, and monitoring.
• You will fuse signals from cameras, microphones, and other sensors to improve scene understanding and instruction following.
• You will create data pipelines, evaluation suites, and simulation experiments that translate directly to real-world performance.
• You will collaborate closely with robotics and cloud engineers to ship features that work in production, not just in the lab.

What You Bring

• You bring strong practical experience with deep learning for perception or multimodal reasoning, and you are comfortable training and fine‑tuning modern models end‑to‑end.
• You understand edge deployment trade‑offs and have hands‑on experience optimizing models for GPU inference on devices such as Jetson.
• You write clean, reliable code in Python and are comfortable with tooling such as PyTorch, CUDA, ONNX, and TensorRT.
• You care about measurement and reliability, and you know how to build datasets, benchmarks, and automated tests that catch regressions before they reach the field.
• You thrive in a fast‑moving environment, and you enjoy turning ideas into working systems that improve every week.