Time since we first saw it
Day 3
Early. Similar jobs usually come down between day 7 and day 76.
From 6,398 similar postings that came down. Coming down is not always a hire.
Copenhagen · Copenhagen, Capital Region, Denmark
Actively hiring, first seen by us 3 days agoBuild with us in Copenhagen This is a full-time role reporting to the CTO. We provide equipment and relocation support. We welcome different backgrounds and paths into machine learning. If the work fits your ambition and you bring substantial relevant experience, apply even if you do not match every item.
Employment type: Full-timeFrom the employer’s posting
Department: EngineeringFrom the employer’s posting
Design and run distributed training, including data pipelines, parallelism and checkpointing
Improve model capabilities through supervised fine-tuning, reinforcement learning and distillation
Find causes of unstable training, poor utilization and unexpected regressions
Build evaluations, including calibrated LLM-as-judge systems, that test whether gains hold
Create and curate training data, including synthetic data pipelines
Designing and running distributed training, including architecture, data pipelines, parallelism, checkpointing, evaluation, and recovery.
Improving model capabilities through supervised fine-tuning, reinforcement learning, and distillation.
Finding the causes of unstable training, inefficient communication, poor utilization, and unexpected regressions.
Building evaluations that expose failure modes and establish whether an improvement holds outside the training distribution, including LLM-as-judge systems you build and calibrate
Creating and curating training data, including synthetic data pipelines that feed real training runs.
You have hands-on experience with the data the models learn from. You have built training-data pipelines, ideally synthetic ones, and stayed close to the training that consumed your data.
You bring strong practical distributed-training experience, and depth in areas such as mixture-of-experts models, reinforcement learning, or distillation. Experience serving open-weight models is a plus. You write code other engineers can build on and make technical trade offs clear.
Actively hiringFirst seen by us 3 days ago, still listed, and not a repost of an earlier posting.
These dates show when we found the posting. Check the employer’s website to confirm it is still accepting applications.
Day 3
Early. Similar jobs usually come down between day 7 and day 76.
From 6,398 similar postings that came down. Coming down is not always a hire.
1 open job
Open jobs by the week we first saw them, last 12 weeks.
Hiring recordAdd your resume to compare the role’s scope, tools and requirements with your experience.
Complete your application on jobs.ashbyhq.com. The employer’s form will show what is required.
Already applied? Track this application
Corti builds AI for critical industries. We started in healthcare, where messy inputs, specialized knowledge, and consequential errors make demanding engineering problems unavoidable. Today we train and post-train models, build agents, and turn research into APIs that developers use to build their products. Our platform supports more than one million production interactions every week.
Our platform and APIs provide the building blocks for specialized intelligence that organizations own:
Models: frontier and open-weight models with reasoning, tool calling, structured output, and embeddings through an OpenAI-compatible interface, plus custom models trained and post-trained on an organization’s own data.
Sovereign compute: GPU capacity on Corti-operated European infrastructure for teams that need their data and workloads to stay in Europe.
Agents: a harness to build, evaluate, and optimize agents that connect to real systems, with planning, tool use, external knowledge, persistent context, human approval, and observable execution.
Speech: live dictation, conversational transcription, and batch audio processing.
Clinical documentation: fact extraction as conversations unfold, and document generation with explicit schemas and controllable synthesis.
Medical coding: code predictions from clinical narratives, linked to supporting passages.
Explore the platform at docs.corti.ai.
You want to train models at the frontier, and you want the science in the open, where other engineers can inspect it. At Corti, you don't have to choose.
As a Senior Machine Learning Engineer in our Applied AI team, you drive large-scale training and post-training alongside researchers and engineers. Your decisions span the training recipe, the evaluation, and the system that serves the result. The calls are consequential, and you stay close enough to the implementation to understand why they succeed or fail.
Designing and running distributed training, including architecture, data pipelines, parallelism, checkpointing, evaluation, and recovery.
Improving model capabilities through supervised fine-tuning, reinforcement learning, and distillation.
Finding the causes of unstable training, inefficient communication, poor utilization, and unexpected regressions.
Developing models for agentic work: choosing tools, using evidence, maintaining context, and completing complex tasks reliably.
Building evaluations that expose failure modes and establish whether an improvement holds outside the training distribution, including LLM-as-judge systems you build and calibrate
Creating and curating training data, including synthetic data pipelines that feed real training runs.
Turning stronger models into practical APIs by improving latency, throughput, memory use, and serving cost.
Which capabilities belong in the weights, and which need tools or retrieval? You help decide, and you build the experiments that establish the answer.
We believe engineers should be able to examine the methods and evidence behind the systems they build on. We publish research, release benchmark resources, and document how our platform works. Read the work behind our clinical APIs: corti.ai/research
You have trained large language models and owned technically complex ML work. You can explain a difficult run in detail: the architecture, the bottleneck, the failed hypothesis, and the evidence that changed your mind.
You bring strong practical distributed-training experience, and depth in areas such as mixture-of-experts models, reinforcement learning, or distillation. Experience serving open-weight models is a plus. You write code other engineers can build on and make technical trade offs clear.
You know post-training methods from the inside.Preference optimisation and reinforcement learning have quirks, reward hacking, runs that look stable until they collapse, and signals that turns out to be noise. You have hit these problems and learned to spot them before they cost you a week.
You have hands-on experience with the data the models learn from. You have built training-data pipelines, ideally synthetic ones, and stayed close to the training that consumed your data.
You have evaluated LLMs in production. You have built and calibrated LLM-as-judge systems to automate that evaluation, and you know when a judge's score means something and when it just flatters the model it is scoring.
You also have the itch to investigate. A new model, paper, or agent framework appears, and you want to understand what actually changed. You build experiments, compare approaches, and test the claims. Perhaps your latest side project involves agents building tools, coordinating with other agents, or failing in a surprisingly instructive way. We want to hear about it.
A relevant advanced degree is welcome; equivalent experience matters just as much.
This is a full-time role reporting to the CTO. We provide equipment and relocation support.
We welcome different backgrounds and paths into machine learning. If the work fits your ambition and you bring substantial relevant experience, apply even if you do not match every item.
Send us something you built, a hard problem you solved, or a result you think deserves to be challenged. A link is enough. Help us shape the models, agents, and infrastructure developers build on next.
________________________________________________________________
We treat all candidates equally, are committed to bringing in top talent from all sectors and countries, and will not discriminate based on race, religion, national origin, gender, sexual orientation, age, marital status, or disability status.
Additionally, we want to remind you that experience comes in many forms, and we're passionate about creating teams with a multitude of perspectives! If you believe your experience is close to what we're looking for but not an exact match, we still hope you consider applying!
New ML Engineer postings at the employers we follow are down 4% over the last 6 weeks compared with the 6 before.
Each bar is one week of new postings. The darkest bars are the last 6 weeks.
Counted from the hiring pages we’ve followed for at least five weeks, so employers we add don’t look like growth. Updated weekly.
We file this job under ML Engineer, by the work the posting describes.
ML engineers put machine learning models into products. They write the code that trains, ships, serves and monitors models so a recommendation or prediction works for real users.
Postings ask for strong Python and software engineering, a framework such as PyTorch, cloud services, and more and more often experience with large language models.
A data scientist often works out which model to build; an ML engineer makes it run reliably in production.
The government has no code for ML engineers, so these figures are for the closest group, computer and information research scientists.
Source: US Bureau of Labor Statistics, Occupational Employment and Wage Statistics (May 2025) and Employment Projections (2025 to 2035). Occupation code 15-1221.
Selected passages from the saved posting. Check the full description for conditions and exceptions.
No pay amount identified in the saved description.
Copenhagen, Capital Region, Denmark
Working pattern and location restrictions need checking in the full posting.
No clear work-authorization passage found. Eligibility is unconfirmed.