Company Description
RoboService solves the "Last Mile" of humanoid robotics. While manufacturers provide the hardware and foundation models, these general-purpose "brains" are not yet ready for the specific, accurate tasks of a production line. We bridge that gap. We don't build foundation models from scratch; we operationalize them. We take state-of-the-art VLA models and fine-tune them for reliability, speed, and precision in real-world manufacturing environments. We are not a research lab; we are shipping code that moves atoms. Launched in July 2025 by repeat founders with successful exits, we are looking for engineers who want to take robots out of the lab and onto the factory floor.
Role Description
We are looking for Senior ML Engineer to bridge the gap between "remote controlled hardware" and "autonomous robots". We work with robotics manufacturers that provide great hardware and a model infrastructure, but are missing the software to fine-tune the models on a specific task. Your primary focus will be establishing training pipelines—from teleoperation data collection to fine-tuning VLA models—and utilizing Reinforcement Learning (RL) to fine-tune complex manipulation and locomotion tasks. You won't be writing papers. You will be:Fine-tuning VLA models (open source and proprietary models from our partners)Managing the Data Pipeline and overseeing the Teleoperation and Imitation Learning activities for data quality. Manage large datasets required for fine-tuning models.Reinforcement Learning (RL): Apply RL techniques to fine-tune robot actions, ensuring robust performance when the robot works on the production line.Utilize high-fidelity simulation environments (NVIDIA Isaac Sim, Omniverse) to train and test agents before deploying them to physical hardware. (Sim2Real)
Qualifications
Education: MS or PhD in Robotics, Computer Science, Artificial Intelligence, or a related field.Experience: 3-5+ years of hands-on experience in robotics, deep learning, or embodied AI.Core ML Skills: Strong proficiency in Python and deep learning frameworks (PyTorch, TensorFlow, or JAX). Experience with Transformers and VLA architectures is highly desirable.Robotics Middleware: Ideally familiarity with ROS2 (Robot Operating System), including navigation, manipulation, and driver integration.Simulation: Proven experience with physics-based simulators, specifically NVIDIA Isaac Sim or similar (Gazebo, MuJoCo).Data Engineering: Ability to manage high quality demonstration datasets, automate data cleaning, and build training pipelines.Pioneering: Experience debugging hardware/software interfaces on non-standard platforms.Experience with C++ is desiredProficiency in English; basic German skills are a plus.
What We Offer
Berlin based: We’re an in‑person team and we work out of our office in downtown Berlin. If you’re not in Berlin, we would like to help you move here and can help with your relocation.Access to Cutting Edge Tech: Work daily on the latest software and hardware that will shape our future.Impact: Work in real projects with manufacturers to bring robots onto manufacturing floors.Startup Environment: We are a fast-moving startup with high growth ambitions.Team of Experts: We are a team of experts where every voice matters. Work with professionals that are leading their field.
#J-18808-Ljbffr
Freelance Robotics & ML Engineer in Berlin Arbeitgeber: Talantir
findIQ ist ein schnell wachsendes Tech-Startup aus der Region Ostwestfalen-Lippe, das sich der digitalen Transformation im industriellen Maschinen- und Anlagendienst verschrieben hat. Als Arbeitgeber bieten wir eine unternehmerische und kollaborative Kultur in einem kleinen, vielfältigen Team, wo Ideen geschätzt werden und die Ausführung zählt. Mit flexiblen Arbeitsmöglichkeiten, individueller Entwicklung und einem attraktiven Gehalt sind wir bestrebt, unseren Mitarbeitern die besten Voraussetzungen zu bieten, um persönlich und beruflich zu wachsen.
