Company interview guide
Siemens Machine Learning Interview Guide (2026)
A short, team-run loop: a screening call, often a test or recorded video, then technical interviews on ML basics, LLMs and your projects.
Siemens's machine learning interview loop has 5 reported stages (Recruiter or HR call, Online test, written exam or recorded video, Technical interview: ML, LLMs and your projects, Coding exercise (some teams), Final round: hiring manager and fit); typical length: Several weeks; candidates mention gaps of one to a few weeks between steps.
Roles covered: ML Engineer, AI Engineer, Gen AI Engineer, Applied Scientist and Research Scientist. The question bank holds 6 questions reported from Siemens interviews.
- Interview process length
- Several weeks; candidates mention gaps of one to a few weeks between steps
- Stages
- 5
- Reports linked to bank questions
- 6
- Bank report dates
- Jul 2023–Sep 2026
- Questions in bank
- 6 (0 free)
- Last reviewed
- October 2026
How the Siemens interview loop works
Apply through the job portal; after CV screening comes a recruiter or HR call, which in a 2026 senior applied-scientist loop also included three behavioral questions. Many teams then use a test: an online or aptitude assessment, a written exam, a timed ML task on a provided dataset, or a recorded one-way video interview. The core is one or two technical interviews covering ML and deep learning basics, LLM post-training or attention for GenAI teams, and a deep dive on past projects; some teams add a short coding exercise and others state there is no coding. A research loop added a manager interview and a closing HR call. The graduate program runs CV screening, an initial interview and an assessment center. This outline rests on few reports, and most do not state a location.
The Siemens interview process, stage by stage
01Recruiter or HR call
Phone or video call with a recruiter or HR; some teams add behavioral questions here
A first conversation about your background and the team. In a research loop it mainly introduced the team; in a 2026 senior applied-scientist loop the recruiter probed LLM post-training experience and asked for an end-to-end project, a failure and a disagreement. The graduate program calls this the initial interview and may hold it by phone, video or in person.
What it tests
- How your experience fits the specific team and product area
- Ownership of a project from start to finish
- How you handle failure and disagreement
- Logistics: location, level, start date
How to prepare
- Prepare three short stories: an end-to-end project, a failure and a disagreement.
- Ask which business unit and team the role sits in; Siemens hires per team and the content follows the team.
- Expect slow replies between steps and follow up through the portal dashboard.
Recruiter or HR call in 4 of 11 ML and AI reports (2022-2026), one with three behavioral questions. The company's graduate program page lists CV screening then an initial interview.
02Online test, written exam or recorded video
Varies by team: a 2-hour online assessment, an aptitude test, a written exam, a timed 2-hour ML task, or a one-way recorded video interview
More than half of our reports mention some kind of test before or during the interviews, but its form differs by team. Interns describe an online or aptitude test and, for one data and AI internship, recorded video answers to written prompts with preparation time. An MLE candidate got a timed recommender task on a provided dataset with a held-out test set, and others sat a written exam or a 45-minute solo test whose results were then discussed in the interview.
What it tests
- Aptitude and general problem solving under time limits
- Building and evaluating a model on a provided dataset in about two hours
- Explaining yourself clearly on camera without an interviewer
- Written answers on ML fundamentals
How to prepare
- Ask the recruiter which format applies; the preparation is very different for each.
- For a timed ML task, get a simple validated baseline submitted early, then improve it.
- Use the practice video interview the portal offers before recording real answers.
- Do not use AI tools during assessments; the company's application rules forbid it unless the test is about AI skills.
Some form of test in 6 of 11 reports: online or aptitude test (2), written exam or solo test (2), timed ML task (1), recorded video (1). The company's FAQ covers video interviews and its graduate page psychometric tests.
03Technical interview: ML, LLMs and your projects
One or two interviews with one or two team members, by phone, video or on site; usually no coding
The round every report describes in some form. Interviewers start from your CV and projects, often spending half an hour or more on a single project, then move to concept questions. Classical ML and deep learning basics come up for most roles; GenAI and LLM teams add attention, Hugging Face tooling and, in the 2026 senior screen, a rapid-fire block on SFT, DPO, PPO, GRPO and LoRA.
What it tests
- Defending the design and tech-stack choices of your past projects
- Neural network basics: ANN vs CNN, gradient descent, optimizers, loss functions
- Classical ML details such as encoding categorical features for decision trees
- Transformer attention and, for GenAI teams, post-training methods and LoRA
- Scenario questions on applying ML in the team's domain
How to prepare
- Know every project on your CV well enough to justify each choice and alternative.
- Revise classical ML and deep learning basics even for LLM roles; candidates who skipped them struggled.
- For GenAI teams, be able to explain what DPO optimizes and why it needs no explicit KL term.
Technical interview in 9 of 11 reports: project deep dive in 5 and concept questions in 6, with LLM topics in 3 since 2024. One research loop had two such rounds and stressed no coding.
04Coding exercise (some teams)
About 20-25 min of coding inside a 60-min technical screen in one report; algorithm-style questions mentioned in another
Coding is not a fixed stage at Siemens. A 2026 applied-scientist screen spent 20 to 25 minutes on a small data pipeline: group variable-length vectors into batches, right-pad each batch and optionally drop the last incomplete one, with corner-case tests. An AI software developer was told to expect algorithm questions, while a research candidate had none.
What it tests
- Clean, deterministic batching and padding logic
- Handling corner cases such as empty input and a short final batch
- Standard data-structure and algorithm questions for software-leaning roles
How to prepare
- Ask whether the technical round includes live coding and in which language.
- Practise writing a small batching and collate function with tests, as in a PyTorch data loader.
Live coding in 2 of 11 reports: a data-pipeline task in a 2026 applied-scientist screen and algorithm questions for an AI software developer in 2025; a 2023 research loop had none.
05Final round: hiring manager and fit
Manager interview, an on-site panel, or an assessment center for the graduate program
Later rounds are the least documented. A research loop ended with a manager interview and a second HR call, and one MLE had a face-to-face day with several people covering soft skills and technical ability. The graduate program ends with an assessment center, virtual or on site, attended by its decision makers.
What it tests
- Motivation for the team and for Siemens
- Collaboration and communication with non-ML colleagues
- Fit with the team's domain and working style
How to prepare
- Have a clear answer for why this team and this business, not just why Siemens.
- Bring questions about the product, data access and deployment on industrial systems.
A later on-site or manager round in 3 of 11 reports (2022-2025), with few details. The company's graduate program page describes a final assessment center.
Questions and themes reported most often
- Deep dive on your own projects and tech-stack choices (5 reports, technical interview: ml, llms and your projects)
- Neural network basics: ANN vs CNN, gradient descent, optimizers (2 reports, technical interview: ml, llms and your projects)
- How attention works in a transformer (1 reports, technical interview: ml, llms and your projects)
- SFT, DPO, PPO and GRPO for LLM post-training (1 reports, technical interview: ml, llms and your projects)
- When to use LoRA and its limitations (1 reports, technical interview: ml, llms and your projects)
- One-hot encoding with decision trees (1 reports, technical interview: ml, llms and your projects)
- Batch and pad variable-length vectors (1 reports, coding exercise (some teams))
- End-to-end ownership, a failure and a disagreement (1 reports, recruiter or hr call)
What Siemens asks about
- ML & deep learning fundamentals35%
- LLM & GenAI concepts25%
- Project deep dive20%
- Coding10%
- Behavioral & fit10%
Siemens interview prep checklist
- Confirmed the team, business unit and the format of every round with the recruiter
- Prepared a project deep dive that defends each design choice
- Revised neural network basics, optimizers and classical ML encodings
- Can explain attention, SFT, DPO and LoRA in plain terms
- Wrote a batching and padding function with corner-case tests
- Rehearsed answers for the practice video interview if one is offered
- Have stories for end-to-end ownership, a failure and a disagreement
Tips
- Find out the business unit and team early; Siemens hires per team and the loop follows the team.
- Ask which test format applies (online assessment, written exam, timed ML task or recorded video) and prepare for that one.
- Rehearse a 30-minute walkthrough of your strongest project, including the alternatives you rejected.
- Revise classical ML and deep learning fundamentals; they appear in most technical rounds.
- For GenAI and LLM teams, prepare attention, fine-tuning, LoRA and preference-optimization methods.
- Expect little or no live coding, but be ready for a short, practical data-handling exercise.
- Follow the company's AI rules: AI help is fine for preparing, not during live interviews or assessments.
- Allow for weeks between steps and check the portal dashboard for updates.
Practice the Siemens loop
Work the stages in order: every reported question for this company, grouped by round, with solutions in the app.
Siemens interview FAQ
How many interview rounds does Siemens have for ML roles?
Candidates report 5 stages: Recruiter or HR call, Online test, written exam or recorded video, Technical interview: ML, LLMs and your projects, Coding exercise (some teams) and Final round: hiring manager and fit. Apply through the job portal; after CV screening comes a recruiter or HR call, which in a 2026 senior applied-scientist loop also included three behavioral questions. Many teams then use a test: an online or aptitude assessment, a written exam, a timed ML task on a provided dataset, or a recorded one-way video interview. The core is one or two technical interviews covering ML and deep learning basics, LLM post-training or attention for GenAI teams, and a deep dive on past projects; some teams add a short coding exercise and others state there is no coding. A research loop added a manager interview and a closing HR call. The graduate program runs CV screening, an initial interview and an assessment center. This outline rests on few reports, and most do not state a location.
How long does the Siemens ML interview process take?
Reported timelines run Several weeks; candidates mention gaps of one to a few weeks between steps.
What does Siemens ask in ML interviews?
By reported question mix: ML & deep learning fundamentals 35%; LLM & GenAI concepts 25%; Project deep dive 20%; Coding 10%; Behavioral & fit 10%. The most repeated themes: Deep dive on your own projects and tech-stack choices; Neural network basics: ANN vs CNN, gradient descent, optimizers; How attention works in a transformer.
How should I prepare for the Siemens ML interview?
Find out the business unit and team early; Siemens hires per team and the loop follows the team. Ask which test format applies (online assessment, written exam, timed ML task or recorded video) and prepare for that one. Rehearse a 30-minute walkthrough of your strongest project, including the alternatives you rejected.
Are there free Siemens interview questions?
LeetML has 6 questions reported from Siemens interviews. The stage sections above link the free practice problems for the topics each round covers; the reported-question set itself is in the app.
Related interview guides
- SAPAn online coding test, an HR screen, then project-driven technical rounds with easy-to-medium coding and a hiring-manager conversation.
- IBMA HackerRank coding and SQL test, then resume-driven ML and GenAI interviews and a manager round; IBM Research runs a research talk instead.
- IntelResume-driven technical conversations on ML fundamentals, light coding, and C++ and hardware awareness for performance teams.
- NVIDIATeam-run loops that mix algorithm coding, deep learning built from scratch and GPU-aware LLM performance.
- QualcommClassic algorithm coding, ML and deep learning fundamentals, and a resume deep dive, with C/C++ and hardware questions on on-device teams.