Meld u aan om toegang te krijgen tot alle onderdelen van onze service.
  • Vacatures zoeken
  • Favorieten
  • Maak een cv
    Nieuw
  • Salarissen
  • Inschrijvingen

[NanoIC topic] Foundation Models, World Models and Continual Learning for Semiconductor Lithography and Metrology: Advancing Machine Learning towards High-NA EUV Process Control

Stage

imec

What will you do?

As patterning moves to high-NA EUV lithography, process windows shrink. Thinner resists produce noisier, lower-contrast SEM images, and stochastic failures (missing contacts, micro-bridges, line breaks) become a dominant yield limiter. Today, computational lithography (OPC, process simulation), metrology and defect inspection are handled as separate problems. Each relies on its own task-specific models, large amounts of labelled data and frequent recalibration. None of them builds a shared, transferable understanding of how design intent, process conditions and measured outcomes relate to each other.

Recent advances in machine learning offer a way to change this. Transformer-based foundation models, Joint-Embedding Predictive Architectures (I-JEPA, V-JEPA), latent world models and continual learning methods can learn from large amounts of unlabelled data, predict outcomes under changing conditions, and adapt to new tasks without retraining from scratch. Semiconductor manufacturing is a demanding testbed for these methods. Its data are noisy, scarce in labels, multi-modal and governed by physics that is only partially known, and it requires nanometre-level accuracy.

This Master’s internship advances these methods in that setting. The semiconductor process flow sets the boundary conditions and evaluation criteria. Depending on the candidate’s profile, the thesis will focus on one or more of the following research directions:

• Self-supervised foundation models for metrology images: pretraining transformer and JEPA-style encoders on large unlabelled SEM/TEM datasets, and studying representations that hold up under low dose, low contrast and anisotropic noise.

• World models of the patterning process: learning action-conditioned latent models that predict the measured outcome from design (layout/GDS) and process conditions (dose, focus, illumination, mask bias). This enables counterfactual “what-if” analysis across the process window.

• Label-free and few-shot defect detection: detecting rare, novel and stochastic defects through prediction error, density modelling or few-shot learning, with calibrated uncertainty.

• Continual learning across layers, tools and nodes: adapting models to new data distributions without catastrophic forgetting, and quantifying how many labelled samples each new task requires.

• Machine learning for computational lithography: fast learned surrogates for lithography simulation, ML-assisted OPC/inverse lithography, and learned mask/process correction, benchmarked against conventional rigorous methods.

• Multi-modal, uncertainty-aware learning: combining metrology images, layout and tool/process data into representations that support reliable process control decisions.

By tackling these challenges, the project aims to advance the state of the art in representation learning, predictive modelling and adaptation under strict noise, data and accuracy constraints. The findings should apply beyond semiconductor manufacturing to other scientific and industrial imaging domains.

The selected intern/student will learn the conventional patterning process flow. They will work collaboratively on developing and applying machine-learning-based algorithms with four goals: 1) reducing computational cost compared to rigorous simulation, 2) reducing tool cycle time and metrology sampling, 3) enabling a predictive process control approach for advanced-node manufacturing, and 4) improving the quality and information content of metrology data.

Machine learning applicability includes:

1. Technical diligence: scope the research question, model choice, compute budget and evaluation protocol to meet the target performance and timeline.

2. Data curation and analysis: collect and analyse image data (SEM/TEM/EDR/…), layout clips and process/tool data. Formulate hypotheses in a feedback loop with domain experts.

3. Model development: build new architectures or substantially improve existing ones (transformers, JEPA, world models, continual learning, generative models) for the chosen task, with a focus on methodological contribution.

4. Rigorous evaluation: benchmark against strong baselines, with ablations, uncertainty quantification and analysis of generalization across data distributions.

5. Dissemination: collaborate on publications targeting top-tier machine learning and computer vision venues (CORE A*/A), such as NeurIPS, ICML, ICLR, CVPR, ICCV and ECCV, and high-impact journals such as IEEE TPAMI, IJCV and TMLR.

What we do for you?

We offer a fully funded internship scholarship (subject to conditions) for a maximum duration of 9 months, contingent on satisfactory progress. The internship focuses on fundamental and innovative research with clear practical applications. You will be supervised towards publishing at top-tier ML/CV venues, with access to real, non-public industrial metrology data rarely available to academic researchers. You will join a young and enthusiastic team of researchers and domain experts in machine learning, lithography and metrology. The Master’s internship position is available immediately.

Who are you?

✓ You hold at least a Bachelor’s degree or are currently pursuing a Master’s degree at a reputable university in Computer Science, Computer Engineering, or Physics with a strong specialisation in advanced machine learning. Coursework in deep learning, computer vision or self-supervised learning is a plus.

✓ You can work independently, demonstrate a strong sense of responsibility, and keep to the timelines and milestones defined by ongoing research projects and conference deadlines. You are genuinely motivated by the research topic and committed to completing a research thesis as part of your university degree programme.

✓ You have a strong research interest in modern deep learning, such as transformers, self-supervised learning, world models, continual learning or generative models, and you are eager to advance the state of the art. Hands-on experience with PyTorch (or JAX) is expected. Experience training large models on GPU clusters is a strong plus.

✓ You have strong programming skills in Python. C/C++ is a plus.

✓ You are motivated to publish at top-tier machine learning and computer vision conferences. Prior publications at such venues are a strong plus.

✓ You are curious about the physical problem behind the data. No prior knowledge of lithography, microscopy or semiconductor processing is required, and the necessary domain background will be provided during the internship.

✓ You have excellent analytical skills for interpreting research results, including negative ones.

✓ You are a team player with strong communication skills.

✓ Your English is fluent, both spoken and written.

Interested?

Send your application, or any questions about this internship, by email to Bappaditya Dey (Het emailadres op chipmakerjobs.com bekijken). Use this subject line: “MS Thesis Application: Foundation Models, World Models and Continual Learning for Semiconductor Lithography and Metrology”.

Applications should include:

(a) SoP (clearly describing your research background and interests, any specific research skills, and which research direction(s) above interest you most; Physics applicants should describe their machine learning coursework and projects)

(b) your professional résumé [max. 2 pages]

(c) your latest degree transcript

(d) GRE/TOEFL score [preferable but not mandatory]

(e) previous publication record and/or links to code repositories (if any)

Our research group website:

Type of internship : PhD internship, Master internship, Bachelor internship

Duration : 6 months, can be extended to 9 months.

Required educational background : Computer Science, Physics

Supervising scientist(s) : For further information or for application, please contact Bappaditya Dey ( Het emailadres op chipmakerjobs.com bekijken ) The reference code for this position is 2026-INT-316 . Mention this reference code in your application. Imec allowance will be provided for students studying at a non-Belgian university.

Applications should include the following information:

resume motivation current study

Incomplete applications will not be considered

Vacature geplaatst op 2 dagen geleden
Soortgelijke banen die u mogelijk interesserenOp basis van de vacature [NanoIC topic] Foundation Models, World Models and Continual Learning for Semiconductor Lithography and Metrology: Advancing Machine Learning towards High-NA EUV Process Control in Leuven
  •  ...As semiconductor devices continue to scale...  ...of high-NA (...  ...(EUV) lithography, defect...  ...and metrology face...  ...in advanced manufacturing...  ...for new...  ...and machine learning...  ...semiconductor process flow...  ...models that...  ...control. • Scalable...  ...to advance the...  ...learn conventional...  ...towards...  ...research topic and... 
    Aanbevolen
    Stage
    Per direct beginnen

    imec

    Louvain
    2 dagen geleden
  •  ...The continuous scaling of semiconductor technology is driving increasingly...  ...dimensions and pitches, enabled by advanced patterning...  ...technologies such as high-NA EUV lithography. As critical dimensions (CDs) continue to shrink and...  ...a growing need for metrology techniques... 
    Aanbevolen
    Stage

    imec

    Louvain
    2 dagen geleden
  •  ...Designing, executing and analysing the experiments for patterning process development using EUV/High NA EUV lithography for memory and...  ...processes and metrology from wafer to electrical...  ...understanding of semiconductor processing and...  ...- Knowledge of advanced lithography and patterning... 
    Aanbevolen
    Fulltime
    Met vast contract

    Engibex

    Louvain
    13 dagen geleden
  •  ...As the world around us becomes...  ...integrated and devices...  ...lithographic processes in which layers...  ...(EUV) light (with...  ...great promise for non-destructive...  .... In metrology it is crucial...  ...number of model parameters...  ...laboratory and advanced patterning...  ...of semiconductor materials.... 
    Aanbevolen
    Stage

    imec

    Louvain
    2 dagen geleden
  •  ...As the semiconductor industry advances beyond...  ...Angstrom era, High-NA and future...  ...-NA EUV lithography are...  ...enablers for manufacturing...  ...etch processes. These...  ...modern machine learning...  ...focus control, and post...  ...generation, metrology...  ...developed to model defect...  ...roadmaps toward High-NA... 
    Aanbevolen

    imec

    Louvain
    2 dagen geleden
  •  ...The continued scaling of semiconductor devices requires advanced characterization techniques...  ...electrical and optical characterization...  ...combines the high spatial...  ...capability for a plethora of...  ...and physical modelling. This will help...  ...characterization techniques. Towards the end of the... 
    Stage

    imec

    Louvain
    2 dagen geleden
  •  ...aspect ratio (AR) and decreasing...  ...great challenges for mechanical stability...  ...stability and defectivity control. High‑AR structures are...  ...and damage during processing steps such as...  ...X‑SEM, and OCD metrology, stability criteria...  ...defined for various process conditions. This... 
    Stage

    imec

    Louvain
    2 dagen geleden
  •  ...The continued scaling of semiconductor technology is leading...  ...dimensions and higher aspect...  ...patterning processes can result...  ...collapse. For example, post...  ...conventional metrology. Many established...  ...with high spatial resolution...  ...dedicated advanced patterning...  ...optical modelling, image... 

    imec

    Louvain
    2 dagen geleden
  •  ...As advanced interconnect patterning...  ...yield, defects, and line...  ...true impact of process variations on...  ...While recent High-NA Direct Metal...  ...Master int...rs for DME |...  ...the design, modeling, and simulation...  ...qualification toward quantitative...  ...will provide a foundation for future electrical... 
    Stage

    imec

    Louvain
    2 dagen geleden
  •  ...by the letter T, and the reflected waves...  ...warpage reconstruction models assume planar,...  ...). Key objectives for this project...  ...edge-correction post-processing algorithm to recover...  ...Strong mathematical foundation in wave...  ...-on cleanroom/lab metrology (SAM, surface profilometry... 
    Stage

    imec

    Louvain
    7 dagen geleden
  •  ...maintaining controlled leakage,...  ...current and...  ...material for such architectures...  ...contacts and processing...  ...develop advanced atomic force...  ...AFM)-based metrology methods (...  ...impacted by process...  ...predictive model linking AFM...  ...selection, force control,...  ...Experience in semiconductor devices,... 
    Stage

    imec

    Louvain
    2 dagen geleden
  •  ...ultraviolet (EUV) lithography has enabled the continued...  ...scaling of semiconductor devices...  ...Moore’s law and enable...  ...the need for ultra-...  ...exceptional control over...  ...fidelity, high quantum...  ...lithography processes has not...  ...scaling toward Hyper-NA...  ...using advanced metrology... 

    imec

    Louvain
    2 dagen geleden
  •  ...Optics for Extreme...  ...Ultraviolet (EUV) Inspection and Metrology...  ...The semiconductor industry is continuously pushing towards smaller...  ...need for high-resolution...  ...inspection, and process...  ...that control the phase...  ...modeling and optical...  ...engineering for advanced...  ...multidisciplinary topic with... 

    imec

    Louvain
    2 dagen geleden
  •  ...We are looking for a highly motivated intern with...  ...artificial intelligence, machine learning, and data-driven process optimization. A...  ...and evaluate AI/ML models that can be integrated...  ...smarter recipe control. We are seeking a proactive...  ...who is eager to learn, enjoys solving... 
    Stage

    imec

    Louvain
    2 dagen geleden
  •  ...Market" in our semiconductor R&D by...  ...parametric extraction and electrical...  .... SQL Agents for our specific...  ...hallucination. Advanced visualization...  ...you will learn Research-...  ...testing. High-Performance computing...  ..., private model inference....  ...Experience processing spatial data... 
    Stage
    Per direct beginnen

    imec

    Louvain
    2 dagen geleden
  •  ...evolution of semiconductor...  ...functionality, and miniaturization...  ...and advanced stacking...  ...die—allow for more...  ...generation of high-...  ...bonding process, akin to...  ...quantitative model allowing...  ...process control and quality...  ...supporting continued innovation...  ...processing and metrology suppliers... 

    imec

    Louvain
    2 dagen geleden
  •  ...limitations of continued lateral...  ...in the advanced...  ...complex and...  ...grand metrology challenge...  ...ultraviolet (EUV) light...  ...technique for...  ...light from high-harmonic...  ...these advances, the advantages...  ...beam control,...  ...characterization of semiconductor devices...  ...modelling of...  ...material and process... 
    Ploegendienst

    imec

    Louvain
    2 dagen geleden
  •  ...is a world-leading...  ...and innovation...  ...in semiconductor manufacturing...  ..., advanced...  ...developed for...  ...for higher-performance...  ...continues to grow...  ...manufacturing processes are...  ..., lithography...  ...and controlling these...  ...and metrology data...  ...Existing machine learning approaches...  ...models and digital...  ...this topic is at... 

    imec

    Louvain
    2 dagen geleden
  •  ...consultant, you contribute to advanced semiconductor patterning process development in a high-tech R&D environment. Your work focuses on EUV and High-NA EUV lithography applications for memory and logic...  ...development, lithography, etch, metrology and electrical testing. By... 

    VeroTech

    Louvain
    6 uur geleden
  •  ...automated, highly...  ...pipeline for advanced micro-...  ...novel machine learning...  ...achieving high detection...  ...and limited...  ...maximize model generalization...  ...by processing raw A-...  ...toward rare defect...  .... In Advances in...  ...scikit-learn). Knowledge...  ...in semiconductor...  ...on lab metrology (SAM).... 
    Stage
    Ploegendienst

    imec

    Louvain
    7 dagen geleden
  •  ...in which n- and p-type devices...  ...particularly attractive for the p-channel...  ...because its high hole mobility...  ...based defect modeling used to...  ...and processing, and guide the...  ...an admittance model, implementing...  ...model to real-world experimental...  ...expertise in semiconductor devices, materials... 
    Stage
    Ploegendienst

    imec

    Louvain
    2 dagen geleden
  •  ...power, performance, and area (PPA) trade-...  ...research for future semiconductor systems requires...  ...tailoring trade-offs towards the envisioned application...  ...will explore, model, and/or optimize...  ..., including both advanced CMOS and beyond-...  ...sequential process, the methodology... 
    Stage

    imec

    Louvain
    2 dagen geleden
  •  ...partners on processing requests, specifications, and deliveries. Project lead for new Low Volume...  ...Definition of process flows based...  ...time and with high quality. Summarize...  ...one of the world’s premier...  ...strong ability to learn and apply...  ...Experience in semiconductor process... 

    imec

    Louvain
    7 dagen geleden
  •  ..., one of the world’s premier research...  ...of the most advanced 3D technology...  ...costs and logic complexity...  ...SoC), offering higher performances...  ...two fully processed wafers with compatible...  ...candidates for the bonding...  ...stress) at high and low...  ..., a lifetime model will be developed... 
    Stage

    imec

    Louvain
    2 dagen geleden
  •  ...Context Semiconductor manufacturing...  ...relies on highly...  ...equipment for lithography, deposition...  ..., thermal processing, cleaning, metrology, and inspection...  ...systems, and control equipment....  ...process equipment...  ...life-cycle model that makes...  ...Skills and Learning Objectives... 
    Stage

    imec

    Louvain
    2 dagen geleden
  •  ...domain. These advanced technologies...  ...sensitivity for subtle...  ...manufacturing defects and, as the...  ..., an airbag controller goes off due...  ...advanced semiconductor technologies...  ...develop new fault models and...  ...the test chip continues to run in its...  ...we hope to learn more about the... 

    imec

    Louvain
    2 dagen geleden
  •  ...Extreme-ultraviolet (EUV) lithography is enabling...  ...than 20 nm for the latest technology...  ...nodes, but the high photon energy...  ..., absorption and etch-resistance...  ...data from a model metal-...  ...EUV doses and process parameters, such...  ...you will then advance the state-of-the... 
    Stage

    imec

    Louvain
    2 dagen geleden
  •  ..., one of the world’s premier research...  ...progress in lithography has...  ...Today’s advanced technologies...  ...cm ² . These high current densities...  ...displacement of atoms and the creation...  ...schemes for copper and advanced...  ...reliable process. Using...  ...measurements, Semiconductor Science and Technology... 
    Stage

    imec

    Louvain
    2 dagen geleden
  •  ...evolution of semiconductor...  ...around (GAA) and...  ...of these high-aspect-ratio...  ...explore advanced topographical...  ...for precise...  ...high-NA EUV, 3D memories...  ...ecosystem of processing and metrology...  ...insights into process optimization...  ...quality control,...  ...supporting continued innovation... 

    imec

    Louvain
    2 dagen geleden
  •  ...is an essential metrology technique in semiconductor manufacturing, enabling high-resolution imaging...  ...device structures and material layers....  ...used to monitor process quality and support...  ...a workflow for SEM cross-section...  ...experience with advanced semiconductor metrology... 
    Stage

    imec

    Louvain
    2 dagen geleden

Wilt u meer vacatures ontvangen?

Abonneer u om vacatures voor [NanoIC topic] Foundation Models, World Models and Continual Learning for Semiconductor Lithography and Metrology: Advancing Machine Learning towards High-NA EUV Process Control te ontvangen. Solliciteer als eerste!