Reinforcement Learning Human Feedback Toolpath Generation in CAD-to-CAM Automation

Organisation/Company KU LEUVEN Research Field Computer science » Informatics Engineering » Computer engineering Engineering » Mechanical engineering Researcher Profile First Stage Researcher (R1) Final date to receive applications 4 Mar 2026 - 23:59 (UTC) Country Belgium Type of Contract Temporary Job Status Full-time Offer Starting Date 1 Apr 2026 Is the job funded through the EU Research Framework Programme? Not funded by a EU programme Reference Number BAP-2026-80 Is the Job related to staff position within a Research Infrastructure?

No

Context

The process of creating a finished part from the CAD model or technical drawing of this part is commonly called the CAD-to-CAM process resulting in a work plan, which is then further used to produce the physical part. Due to the high complexity, this process is time‑consuming and typically performed by an experienced operator/engineer. It includes the planning of different processing steps, e.g., cutting methods, machining patterns, and sequences.

Selecting the tool, clamping strategy, cutting patterns and parameters requires the availability of experienced operators, which is a challenge today. Especially in the case of complex parts, CAD-to-CAM is a tedious manual process and often results in suboptimal designs, both in terms of material removal rate, path length, idle time, tool change frequency, and quality. A traditional Computer‑Aided Design & Manufacturing (CAD-to‑CAM) workflow assists this process but is limited to validating the outcome of the operator decisions.

PhD research project

The objective of this PhD is to contribute to the automated generation of work plans, for both existing and new parts to be produced in one clamping operation, in order to support the operators and indirectly the production planners/engineers. To this end, it will be investigated how a novel machine learning‑based methodology leveraging reinforcement learning with human feedback and multi‑objective optimisation can be realised to generate new and even improve existing work plans, to semi‑automate the CAD‑to‑CAM process.

To this end, 3 main data sources will be leveraged: (i) the product design data (CAD files), (ii) historical work plans and process information, including machining parameters, constraints, tools, and strategies (i.e., the selection of processing types and their sequence), and (iii) historic machining quality data. The focus is on the mainstream machining processes namely turning, milling, and drilling.

This PhD position is part of the Flanders Make Strategic Basic Research project AutoCAM, which intends to largely automate the generation of work plans, for both existing and new parts to be produced in one clamping operation, to support the operators and indirectly the production planners/engineers in industry. It offers the opportunity to perform the research in close collaboration with leading industry partners.

We are seeking a highly motivated, enthusiastic, passionate, and communicative researcher, with a proactive and creative attitude who is eager to explore innovative solutions. If you recognize yourself in the story below, then you have the profile that fits the project and the research group:

• I have a Master's degree in Computer Science, Artificial Intelligence, Electrical Engineering, Mechanical Engineeringor a related field and performed above average in comparison to my peers.
• I am proficient in written and spoken English.
• During my courses or prior professional activities, I have gathered experience with machine/deep learning, and can demonstrate a strong affinity with these fields. Prior experience with reinforcement learning, multi‑objective optimisation and/or the CAD/CAM process is a plus.
• I am proficient in Python and am familiar with data science and machine/deep learning toolkits.
• As a PhD researcher at KU Leuven, I perform research in a structured and scientifically sound manner. I read technical papers, understand the nuances between different theories and implement and improve methodologies myself.
• Based on interactions and discussions with my supervisors and the colleagues in my team, I set up and update a plan…