Sr. Computer Vision Engineer / Hyderabad

Alchemy Hr Services; Hyderabad; Hybrid; INR 1,500,000 - 2,100,000; Full time

Job summary

Alchemy Hr Services is seeking an Industrial Computer Vision Engineer in Hyderabad to design, deploy, and operate production-grade CV systems for industrial safety and quality use cases. The role is hybrid (4 days office, 1 day WFH) and requires hands-on expertise from model development to production deployment.

The candidate will own CV solutions end-to-end, integrate with plant systems via REST APIs, and ensure robust performance under challenging industrial conditions.

Qualifications

• Hands-on, production-level CV/AI experience needed.
• Strong Python/C++ programming and CV fundamentals.
• Experience with edge deployment and production ML systems.

Responsibilities

• Design and train CV models for industrial use cases, including detection and quality inspection.
• Deploy and optimize models on edge hardware with TensorRT/ONNX/OpenVINO.
• Own full lifecycle of deployed CV solutions with API integrations.
• Embed privacy and governance by design and maintain runbooks.

Skills

Python

C++

OpenCV

PyTorch

TensorFlow

REST API

Cloud platforms

Software engineering

Edge deployment

Education

Bachelor's or Master's in CS/Engineering

Tools

TensorRT

ONNX Runtime

OpenVINO

Job description

Job Role: Industrial Computer Vision Engineer

Location: Hyderabad, India

Employment Type: Full-Time

Experience Required: 6-10 years

Work Mode: Hybrid (4 days office & 1 day WFH)

Role Summary

We are looking for an Industrial Computer Vision Engineer to design, deploy, and operate production-grade computer vision systems for industrial safety and/or quality use cases within live plant environments.

This is a hands-on engineering role focused on edge-deployed CV not academic research or generic data science. You will typically build deep expertise in one primary domain (safety monitoring or visual quality inspection) while maintaining working familiarity with the other and will own your solutions from model development through optimized, monitored production deployment.

Reliability matters as much as accuracy here: false positives in safety systems erode trust and disrupt operations, while unstable outputs in quality systems can halt production lines or trigger unnecessary rework.

Key Responsibilities

Model Development

• Design and train computer vision models for industrial use cases object/person detection, zone-based intrusion detection, visual quality inspection (defects, anomalies, presence/absence), and process-compliance verification.
• Combine ML inference with deterministic, rule-based logic for safety enforcement, quality gates, and auditable decision-making.
• Validate models against real plant footage, accounting for dust, smoke, low or uneven lighting, occlusion, vibration, and fixed industrial camera viewpoints.

Edge Deployment & Optimization

• Deploy and optimize models on edge hardware using TensorRT, ONNX Runtime, or OpenVINO, targeting stable FPS, predictable latency, and controlled compute/memory footprint.
• Apply model compression and optimization techniques, balancing accuracy against throughput and false-alarm sensitivity.
• Design pipelines that degrade gracefully under camera dropouts, connectivity loss, or partial hardware failure.

Production Ownership & Integration

• Own the full lifecycle of deployed CV solutions from training to monitored, maintained production systems.
• Build REST API integrations connecting CV pipelines to plant systems and downstream applications.
• Collaborate with Data Scientists, Software Engineers, Cloud/Platform teams, and Manufacturing Operations to align technical solutions with plant-floor realities.

Governance & Responsible AI

• Embed privacy and workforce-protection safeguards by design, including face/identity anonymization and controlled video retention where required.
• Maintain clear technical documentation covering model behavior, failure modes, and operational runbooks.

Mandatory Skills & Qualifications

Candidates should be able to demonstrate hands-on, production-level experience not just familiarity across the areas below.

• Strong programming skills in Python and/or C++.
• Solid computer vision fundamentals: OpenCV, object detection, object tracking, image segmentation, and video analytics.
• Hands-on deep learning experience with PyTorch or TensorFlow, including model training and fine-tuning.
• Practical edge AI deployment experience using at least one of: TensorRT, ONNX Runtime, or OpenVINO.
• Demonstrated model optimization and compression work (quantization, pruning, latency/throughput tuning).
• At least one real production deployment of a CV system (industrial or otherwise) not only prototypes or research pipelines.
• Working understanding of real-time and/or distributed AI inference systems.
• Experience building and consuming REST APIs for system integration.
• Working exposure to at least one major cloud platform (AWS, Azure, or GCP).
• Solid software engineering practices version control, testing, code review, and system design fundamentals.

Good to Have Skills

• Experience in industrial manufacturing, quality inspection, or safety monitoring systems.
• Exposure to Vision Language Models (VLMs), multimodal AI, or generative AI.
• MLOps experience, including AWS SageMaker, AWS Greengrass, or Azure ML.
• Experience with privacy-preserving CV architectures (anonymization, output abstraction).
• Familiarity with edge camera hardware and industrial interfaces GigE Vision, OPC-UA, MQTT, or Modbus.
• Experience mentoring engineers or informally leading technical initiatives.
• Strong technical writing and stakeholder communication skills.

Required Experience

6-10 years of overall experience in computer vision / applied AI engineering, with at least 2-3 years involving edge deployment or production ML systems (industrial or non-industrial).

• 6-7 years: candidates should show strong hands-on CV/DL skills plus demonstrable exposure to at least one production or edge deployment — not model-development experience alone.
• 8-10 years: candidates are expected to bring multiple end-to-end deployments, direct experience with performance tuning under real operating constraints, and readiness to mentor junior engineers.

Education Requirements

Bachelor’s or master’s degree in computer science, Electronics, Robotics, or a related engineering discipline is preferred — OR equivalent practical experience with demonstrable production computer vision deployments in lieu of a formal degree.