Senior Computer Vision Engineer

MonoEdge; Pune District; On-site; INR 3,000,000 - 6,000,000; Full time

Job summary

MonoEdge in Pune, India is building an advanced perception stack for the Indian manufacturing sector. You will own defect detection, surface inspection, and multi-camera integration on production lines, designing datasets and models from scratch.

You will ship edge-deployed perception models and integrate with PLCs for real-time operation, collaborating with a senior team to deliver measurable plant-floor improvements.

Qualifications

• 4+ years of computer vision experience with at least 2 years building production systems — shipped code, not just research notebooks.
• Deep learning fundamentals: CNNs, segmentation (UNet, Mask R‑CNN, and similar), object detection, classification. PyTorch preferred.
• OpenCV, geometric calibration, multi-camera setups, lighting design. Familiarity with Basler, FLIR, Hikrobot or similar sensors; GigE/USB3 Vision, exposure and gain tuning, lens selection.
• Strong Python: clean code, version control, testing, packaging. Can ship to a small team without supervision.
• Pragmatism about data: you know that 80% of CV work is data, and you have the patience and discipline for it.

Responsibilities

• Specify and deploy camera and capture infrastructure at our first partner plant.
• Build datasets from scratch: capture, label, version, manage. Define what

Skills

Computer Vision

Deep Learning

Python

CNNs

Object Detection

UNet

OpenCV

Industrial Cameras

Data Labeling

Tools

GigE Vision

USB3 Vision

Calibration

Version Control

Job description

MonoEdge develops an advanced intelligence layer for the Indian mid-market manufacturing sector. We provide data-driven insights and strategic recommendations to plant leadership and operations supervisors across diverse linguistic contexts, including English, Hindi, and Marathi.

As a technically-driven, bootstrapped organization, we are defining a new category in industrial optimization.

What you'll work on

• Defect detection and surface inspection on production lines: classifying, localising, and characterising material defects from imagery.
• Geometric measurement and dimensional analysis of moving parts and continuous processes.
• Multi-camera systems for production lines: camera selection, lensing, lighting, mounting, calibration, and synchronization with PLC-driven processes.
• Model development — segmentation, detection, classification — trained from datasets you will help create and curate from scratch.

What you'll do

• Specify and deploy camera and capture infrastructure at our first partner plant. Hands-on, industrial, not a desk job.
• Build datasets from scratch: capture, label, version, manage. Define what \"good labels\" mean for an industrial problem where no public dataset exists.
• Build and train perception models (segmentation, detection, classification) and iterate them based on production validation.
• Partner with the data scientist on the team to connect perception outputs into downstream decision models.
• Productize the capture and inference stack so plant N+1 deploys in weeks, not months.
• Take systems from offline / advisory mode into real-time, edge-deployed, closed-loop integration with plant PLCs.

What you bring

• 4+ years of computer vision experience with at least 2 years building production systems — shipped code, not just research notebooks.
• Deep learning fundamentals: CNNs, segmentation (UNet, Mask R‑CNN, and similar), object detection, classification. PyTorch preferred. Comfortable training from scratch and fine-tuning pretrained backbones.
• Classical CV: OpenCV, geometric calibration, multi-camera setups, lighting design. You don't reach for a neural network when a Hough transform will do.
• Industrial camera experience: Basler, FLIR, Hikrobot, or similar. Comfortable with GigE / USB3 Vision, exposure and gain tuning, lens selection.
• Strong Python: clean code, version control, testing, packaging. Can ship to a small team without supervision.
• Pragmatism about data: you know that 80% of CV work is data, and you have the patience and discipline for it.

Nice to have

• Industrial or manufacturing CV exposure — quality inspection, defect detection, process monitoring, robotics.
• PLC / automation awareness — you don't need to write ladder logic, but knowing what OPC UA / Modbus tags are, and how to integrate with them, helps.
• Thermal or near-IR imaging experience.
• Experience designing capture systems for hot, dusty, or otherwise hostile environments.

How we work

Working directly with the technical Founder, you will maintain end-to-end ownership of the computer vision stack. This position demands a senior professional capable of making high-level architectural decisions, selecting optimal model frameworks, and managing complex technical trade-offs with full autonomy.

Our culture prioritizes efficiency, direct communication, and tangible operational outcomes. We define success through measurable plant-floor improvements, such as yield optimization and significant cost reduction, ensuring our technical solutions deliver clear business value.