Rushikonda IT SEZ in Visakhapatnam is where this Senior Computer Vision Engineer role is based: a full-time, on-site position paying up to ₹1,77,500 per month.
It's one of the computer vision engineer jobs for experienced professionals in Visakhapatnam rather than a first role in the field. Whoever takes it will be trusted to own a vision model from early experimentation through to something actually running in production, not just handed a narrow slice of an existing pipeline. The team sits within a broader data science and machine learning group, working on a handful of vision-driven products rather than one single application.
Background and experience
A bachelor's degree covers the education requirement, and a quantitative or computing background is typical, though what tends to matter more at this level is a track record of applied work rather than the specific degree title on a transcript. A portfolio of shipped machine learning projects, or published research in the space, carries real weight during screening, more so than it would for a more junior opening.
Candidates should have 66 months of experience where the actual day-to-day work involved writing Python code to train a model and then getting it running somewhere real, with a meaningful chunk of that time spent specifically on vision problems rather than general-purpose ML work. Time spent on a single long-running vision project counts for more here than a scattered set of short experiments across different domains.
Skills
- Strong Python skills for building, training, and deploying models
- Hands-on OpenCV experience for image and video processing tasks
- Time spent training and tuning models inside at least one of the major neural network libraries
- Real image processing background, including the preprocessing and cleanup work that happens before a model ever sees the data, since raw footage rarely arrives in a state a model can use directly
Good to have
Experience with object detection or segmentation architectures such as YOLO or Detectron2 is a real plus here, and so is any exposure to optimizing models for edge deployment using something like TensorRT or ONNX. A background in camera calibration or basic 3D vision work helps on certain projects, though it's far from universal across the kinds of work this team takes on. None of these bonus items are required, but a candidate with hands-on time in even one of them usually needs less ramp-up on their first project.
What the work involves
The scope covers the full lifecycle of a vision model, not just the training step in the middle.
- Build and train computer vision models for tasks like object detection, classification, or segmentation
- Work directly with image and video datasets, handling the preprocessing, augmentation, and quality issues that show up in raw visual data
- Deploy trained models into production systems where they need to run reliably and fast enough for whatever they're supporting
- Track model performance after deployment and catch accuracy drift before it turns into a real problem downstream
That last point matters more than it sounds like it should. A vision model that performed well during training can quietly degrade once real-world lighting, camera angles, or image quality start drifting from what it was originally trained on, and catching that early is often the difference between a small fix and a much bigger one. Waiting until users start complaining is usually the most expensive way to find out something's wrong, and by then the fix often costs more engineering time than it would have earlier.
Where you'd be working
The role sits in Rushikonda IT SEZ, a coastal stretch of Visakhapatnam that's grown into a real hub for tech employers over the past several years. Naukri Mitra lists computer vision engineer jobs in Rushikonda IT SEZ, Visakhapatnam fairly regularly, since the SEZ has attracted a steady mix of product companies and research-heavy teams looking for a location outside the more saturated metro cities. The coastal setting is a genuine draw for some candidates weighing this against a similar role in a more landlocked city, even if it's a minor factor next to the actual work.
Pay and benefits
Pay for this role goes up to ₹1,77,500 per month, toward the higher end of what a computer vision engineer's salary in Visakhapatnam looks like at the senior level. The package includes health insurance, paid time off, and performance-linked bonuses tied to project outcomes.
Since the position is on-site, relocation assistance and accommodation support are also available for anyone moving to Visakhapatnam specifically for the job. That support tends to matter given how narrow the pool of experienced vision engineers is outside a handful of larger metro cities.
Team and pace
Work tends to alternate between long research-heavy stretches, where progress is measured in small accuracy gains over several weeks, and tighter deployment pushes once a model is ready to move from a notebook into a real system. Both modes ask for a different kind of patience, and people who only enjoy one tend to burn out on the other side of the cycle eventually, which is worth being honest about going in.
Most senior engineers on this team came up through a couple of years of more general machine learning work before specializing in vision, since the underlying model-training instincts transfer even when the data type changes. From here, some move toward a research-heavy track focused on new model architectures, while others move into a role spanning multiple product teams, applying vision work more broadly.
Applying
Come ready to walk through a vision model you've personally taken from an early prototype to something deployed, including a moment where real-world data behaved differently than the training set predicted. A portfolio link or a couple of relevant papers matter more here than a long list of frameworks on a resume. It also helps to mention which specific vision task you've spent the most time on, whether that's detection, segmentation, or something else entirely, since that shapes which parts of the interview will go deepest. Applications are reviewed on a rolling basis, and shortlisted candidates hear back directly for the next round.
