Machine learning engineer
The ML engineer builds and ships models as software. Day to day that means training pipelines, feature code, evaluation harnesses, serving infrastructure and the glue between them. In many teams the ML engineer also owns the model in production: monitoring drift, retraining, and fixing it when it degrades.
Listings typically ask for strong Python, one deep learning framework (PyTorch is the common one), SQL, experience with cloud infrastructure and some evidence you have taken a model from a notebook to a service. It is the largest category on TixelJobs by a wide margin, 12,399 live listings, and 29% of apply clicks since February 2026 went to it.
Applied scientist
Applied scientist is a title some companies use for a role that sits between ML engineer and research scientist. The applied scientist is expected to read papers and adapt methods to a product problem, run rigorous experiments, and hand a working approach to engineers or ship it with them. The emphasis is on choosing and validating the method rather than operating it.
Requirements lean toward a master's or PhD, publications or a comparable track record, and depth in one area such as recommendation, ranking, NLP or vision. On TixelJobs these roles are usually filed under research scientist or under the specialty hub they belong to, so search both.
Research scientist
A research scientist advances the method itself. The output is papers, new models, benchmarks and prototypes; production readiness is someone else's job. Interviews test depth of understanding, the ability to design experiments and a record of original work.
A PhD or equivalent is the norm, and the bar for publications is high at the labs everyone has heard of. There are 1,526 research scientist listings live on TixelJobs, which makes it a small category relative to engineering, and the competition per role is stiffer.
Data scientist
The data scientist answers questions with data: what happened, why, and what will happen if we change this. The tools are SQL, Python or R, statistics, experimentation and dashboards. Some data scientists build models, but the models are usually there to inform a decision rather than to run inside a product.
The title has narrowed over the past few years as companies split it into analytics roles and ML roles. That shows up in the numbers: 1,414 data scientist listings on TixelJobs against 12,399 for ML engineer. If you like modeling more than reporting, search ML engineer too.
MLOps and ML platform
MLOps engineers build the infrastructure that ML engineers use: training clusters, feature stores, model registries, deployment pipelines, GPU scheduling and monitoring. The role is closer to platform engineering than to modeling, and listings ask for Kubernetes, Terraform, CI/CD, cloud services and enough ML knowledge to understand what the pipeline is for.
It is the fifth largest category on TixelJobs at 2,086 listings. If you come from DevOps or backend infrastructure, it is the most direct route into AI work.
The specialty and adjacent titles
Several categories are defined by the problem rather than the function. NLP and LLM roles (2,723 on TixelJobs) cover language models, retrieval, fine-tuning and evaluation. Computer vision roles (1,106) cover image and video. Robotics roles (2,766) combine perception, control and hardware. Each of these is usually an ML engineer or research scientist job with a domain attached, so search the specialty hub and the function hub both.
Data engineer is the big adjacent title with 6,549 listings. Data engineers build the pipelines and warehouses that everything else depends on, and the role is a common way into ML engineering. AI annotation (1,083 listings) covers data labeling and quality work, often at entry level. AI product manager (290) is a small category for people who define what gets built.
How the titles overlap, and how to search
The boundaries between these titles are set by each company's org chart, not by the industry. A few rules of thumb help when you are reading listings:
- A startup ML engineer often does research, engineering and MLOps at once. A large company splits the same work across three teams.
- Read the responsibilities section before the title. If half of it is about experiments and papers, it is a research role whatever the header says.
- Search by skill as well as title. A query for PyTorch and inference will surface ML engineer, MLOps and applied scientist roles together.
- Level changes the meaning. A senior data scientist at one company does what a staff ML engineer does at another.
Where to look next
Each title above has its own hub on TixelJobs: the ML engineer jobs hub, the data scientist jobs hub, the research scientist jobs hub, the MLOps jobs hub, the NLP jobs hub and the computer vision jobs hub. The remote AI jobs page cuts across all of them, and the full search lets you filter any title by level and country. Browsing is free. Membership is $9 a year or $4.90 every three months and unlocks the apply link and the full description on every job. Cancel any time from your billing page.