A company at Technopark, Kazhakkoottam, in Thiruvananthapuram is hiring an MLOps Engineer for a full-time, on-site role paying up to ₹1,07,500 a month. The role expects three years of hands-on experience deploying machine learning models, and it sits at the point where a data science team's work either makes it into production or stalls before it reaches a real user. The company's data science team has grown faster than its deployment infrastructure over the past year, so this hire is meant to close that specific gap.
What the work looks like
Data scientists on this team build models. Your job is to make sure those models run reliably once they leave a notebook, which is a much larger undertaking than most people expect.
- Build and maintain ML pipelines that take a model from training through to production deployment
- Containerize models using Docker and orchestrate them with Kubernetes
- Set up CI/CD workflows so model updates ship consistently rather than through manual, error-prone steps
- Monitor deployed models for performance drift, since a model that worked well at launch can quietly degrade as real-world data shifts underneath it
A recent example: a churn prediction model that performed well in testing started producing noticeably worse recommendations about four months after launch, and tracing it back took real digging. The input data distribution had shifted after a product change nobody on the ML team had been looped in on. Catching that kind of drift before it does real damage to a business decision is core to this role, not an occasional side task. It's rarely obvious from the outside that a model has quietly started performing worse, which is exactly why proper monitoring has to be built in from the start rather than added after something goes wrong.
Skills that matter here
- Python, comfortable enough to work across both data science code and infrastructure scripting
- ML pipeline tools, whether that's a specific framework or a custom setup built from smaller pieces
- Docker, for packaging models and their dependencies consistently
- Kubernetes, for running and scaling those containers in production
- CI/CD, so deployments are repeatable rather than a manual process someone has to remember each time correctly
Beyond the core list, experience with a tool like MLflow or Kubeflow for experiment tracking and pipeline orchestration is a genuine plus. Familiarity with a cloud ML platform, whether that's SageMaker, Vertex AI, or Azure ML, helps too, since much of the infrastructure work happens in a managed cloud environment rather than on bare servers. Some exposure to monitoring tools built specifically for model drift, rather than general infrastructure monitoring, rounds this out nicely. None of these extras are required, but a candidate who's already touched one or two will spend less time getting oriented and more time contributing to the pipelines that need attention right now.
Education and experience
A bachelor's degree is required, generally in computer science, statistics, or another quantitative field, though what matters more in practice is demonstrated experience building and deploying real models rather than the specific degree title. Three years of hands-on experience is the baseline for this seat, ideally including real ownership of getting at least one model from a research environment into a live production system. Candidates coming purely from a data science background, without deployment experience, will likely need to show genuine interest in the infrastructure side before the team considers them a strong fit. In the interview, the hiring team looks for whether a candidate can explain why a specific production issue happened, not just recite a list of MLOps tools they've heard of.
Team and how the work fits together
This role sits between the data science team and the broader engineering organization, translating models built by one group into systems the other group can actually run reliably. One of the data scientists on this team joined through Naukri Mitra last year, and the working relationship between her and whoever takes this role will likely be one of the closer day-to-day collaborations. Expect regular back-and-forth about what a model actually needs to run well in production, since data scientists and infrastructure engineers don't always start from the same assumptions. A model that's mathematically excellent but computationally expensive to run at scale is a genuinely common source of friction, and part of this role is helping the data science team understand those trade-offs earlier in the process rather than after a model is already built.
Some weeks lean heavily toward building new pipeline infrastructure, others toward keeping existing deployments healthy and responding when a model's performance starts trending the wrong way. Both are a normal part of the role, and the mix shifts depending on which models are newly launched versus already stable.
Compensation and benefits
- Salary up to ₹1,07,500 per month
- Health insurance
- Paid time off
- Provident fund contributions
- Relocation assistance and accommodation support for candidates moving to Thiruvananthapuram
Anyone comparing MLOps engineer salary in Thiruvananthapuram per month figures across Technopark will find this reasonably competitive for the required experience level, especially with accommodation support factored into total compensation. Provident fund contributions follow the standard statutory structure, which is worth knowing when weighing total compensation across different offers.
Applying
This is an on-site role in Technopark, Kazhakkoottam, so relocation or a reliable local commute is part of the arrangement. Send a resume that details specific models you've helped deploy, including the tools involved and any production issues you've resolved after launch. Candidates browsing MLOps engineer job openings in Thiruvananthapuram should expect a technical round covering pipeline design and a past deployment challenge, followed by a shorter conversation with the data science team. Most candidates hear back within a week of the final round. Come prepared to walk through one deployment that didn't go smoothly the first time, since that tends to reveal more than a description of a clean, uneventful launch.
