Machine Learning Engineer, fully remote, full-time, $142,000 a year, open worldwide.
Responsibilities
- Design, build, and deploy scalable ML systems that hold up under real production load
- Optimize model performance and the infrastructure it runs on, not just the model itself
- Work directly with data scientists to move models from prototype into production
A recommendation model for a recent project scored well on every offline metric, yet conversion still dropped after launch. The cause turned out to be a subtle mismatch between the training pipeline and the production feature pipeline: one input was being computed slightly differently in each. Catching that kind of gap, where a model looks correct in isolation but behaves differently once it's actually serving traffic, is a core part of this job.
Infrastructure work sits right alongside the modeling. Feature stores need maintaining, serving latency needs watching, and retraining pipelines need to keep running reliably as data volume grows. None of that shows up in a typical portfolio, but a model with great accuracy that takes eight seconds to return a prediction isn't actually useful in most production settings.
Skills
- Python
- Machine learning frameworks
- Data structures and algorithms
- SQL
- Cloud platforms
- Model deployment
- Software engineering practices
The software engineering fundamentals matter as much as the ML-specific skills here. A model that's technically sound but wrapped in code nobody else on the team can maintain creates more work down the line than it saves. Version control discipline, testing, and clear code structure are expected to be at the same level as in any production software role.
SQL comes up more often than people expect for an ML role, mostly for pulling and validating training data before it ever touches a model. Someone who's comfortable writing a clean, efficient query will spend less time waiting on data and more time on the actual modeling work.
Experience and background
This role requires 30 months of demonstrated experience building and deploying machine learning systems, meaning real production work rather than research projects or coursework alone. The formal education requirement is a bachelor's degree, generally in computer science, engineering, or a closely related field. A fair number of engineers at this level also hold a master's degree, though it isn't required, and a strong production track record carries more weight in review than an advanced degree without applied experience behind it.
Candidates coming from a pure research background sometimes underestimate how much of this role is software engineering rather than model experimentation. Someone who's spent most of their career fine-tuning models in a notebook, without much experience getting them into a live system, will find the production side of the job a real adjustment.
The 30 months doesn't need to come from a single job titled "Machine Learning Engineer." A software engineer who spent that time increasingly focused on ML infrastructure, or a data scientist who moved toward the deployment side of projects, both fit the experience bar as long as the production ML work is genuinely there.
Pay and benefits
The role pays $142,000 a year. It comes with health insurance, paid time off, and a 401(k) match, along with a real budget for training and conferences, since keeping up with a field that moves this fast takes ongoing investment rather than a one-time onboarding course.
Naukri Mitra is coordinating hiring for this opening, and the training budget has historically gone toward things like specialized ML certifications, industry conference attendance, and occasionally a focused course on a specific framework the team's adopting.
For context on remote machine learning engineer salaries at this experience level, this offer sits comfortably above the median, reflecting both the production-deployment focus of the role and the demand for engineers who can operate on both the ML and infrastructure sides.
How the work happens
This team splits its time fairly evenly between building new models and maintaining what's already in production, which is a different balance from many machine learning engineer jobs worldwide that lean heavily toward one or the other. Maintenance here isn't an afterthought; a model that quietly degrades over time because nobody's closely monitoring its performance metrics is treated as a real problem, not background noise.
Collaboration with data scientists runs both directions. Sometimes a data scientist hands off a working prototype that needs to be made production-ready. Other times, engineering constraints discovered during deployment send a model back for rework before it ships, and that back-and-forth is a normal part of how a model gets from idea to something users actually interact with.
Deployment cadence varies by project. Some models ship and stay largely unchanged for months. Others get retrained and redeployed on a regular schedule as new data comes in, and knowing which situation calls for which approach is part of the judgment this role expects.
New hires typically spend their first few weeks getting familiar with the existing model inventory and the infrastructure supporting it before taking ownership of anything new. Every ML system accumulates its own quirks over time, and understanding those quirks tends to matter more early on than jumping straight into a new project.
Applying
Send a resume along with a short description of a machine learning system you took from prototype to production, including one problem that only showed up once it was live. Interviews include a technical round covering both ML concepts and software engineering practices, followed by a system design conversation focused on deploying and maintaining models at scale.
