AI & Computer Vision Engineer

Techasoft Pvt. Ltd.; Bengaluru; On-site; INR 1,800,000 - 2,800,000; Full time

Job summary

Techasoft Pvt. Ltd. is seeking an AI and Computer Vision Engineer to design, train, and deploy vision models that operate on factory floors and enterprise systems.

You will integrate camera data with MES, VIN, and BoM APIs, and implement edge inference with robust dashboards for audits. The role requires 2+ years of hands-on Python development, experience with OpenCV and a DL framework (TensorFlow or PyTorch), and strong English communication to explain results to non-technical clients.

Qualifications

• 2+ years of hands-on Python development.

Responsibilities

• Build and train deep learning models for defect detection, object detection and classification on industrial camera images.
• Write OpenCV pipelines for preprocessing, contour detection and measurement on conveyor and inspection-station feeds.
• Deploy models for edge inference and keep them updated from a central place.
• Connect vision output to plant systems such as MES, VIN and BoM APIs, and to rejection mechanisms like air jets or robotic arms.
• Build anomaly detection and risk scoring models on event streams with clear confidence scores for human reviewers.
• Produce dashboards, audit records and image evidence that clients can trust in audits.
• Track accuracy, false rejects and missed defects, and retrain when production conditions change.
• Write clean, documented Python code and take part in design and code reviews.

Skills

Python development

Deep learning

CNN detection

Dataset labelling

Model evaluation

REST APIs

Git & Linux

English communication

Tools

OpenCV

TensorFlow

PyTorch

Docker

Linux

PostgreSQL/MySQL/MongoDB

REST APIs

Job description

Job Description

About Techasoft and the work you will join

Techasoft is a Bangalore-based software company, founded in 2016, with a growing focus on enterprise AI and computer vision work. You will join the team that builds systems like these three:

• Bad biscuit detection and rejection : deep learning plus OpenCV contour detection on high-speed industrial cameras, with automatic rejection of defective units. It reached 94% defect detection accuracy.
• Mahindra vehicle pre-delivery inspection : multi-camera checks at inspection stations, matched to each vehicle's variant and Bill of Materials through VIN and MES APIs, with edge AI inference, automatic pass or hold decisions and a timestamped audit record.
• AI employee monitoring and compliance : behavioural analytics on Kafka and Spark, with TensorFlow models for baselines, anomaly detection and risk scoring, and a React dashboard for human review.

Job: AI and Computer Vision Engineer

Bangalore | Full Time | Starts immediately | 2+ years of experience

You will design, train and deploy vision and analytics models that run on real factory floors and enterprise systems, from camera capture to the final pass, reject or alert decision.

Responsibilities

• Build and train deep learning models for defect detection, object detection and classification on industrial camera images.
• Write OpenCV pipelines for preprocessing, contour detection and measurement on conveyor and inspection-station feeds.
• Deploy models for edge inference and keep them updated from a central place.
• Connect vision output to plant systems such as MES, VIN and BoM APIs, and to rejection mechanisms like air jets or robotic arms.
• Build anomaly detection and risk scoring models on event streams, with clear confidence scores for human reviewers.
• Produce dashboards, audit records and image evidence that clients can trust in audits.
• Track accuracy, false rejects and missed defects, and retrain when production conditions change.
• Write clean, documented Python code and take part in design and code reviews.

Requirements

• 2+ years of hands-on Python development.
• Working experience with OpenCV and at least one deep learning framework (TensorFlow or PyTorch).
• Practical knowledge of CNN-based classification and object detection, and of dataset labelling and augmentation.
• Understanding of model evaluation: precision, recall and accuracy in a production setting.
• Experience with REST APIs and databases (PostgreSQL, MySQL or MongoDB).
• Familiarity with Git, Docker and Linux.
• Clear written and spoken English, and the ability to explain results to non-technical clients.

Good to have

• Streaming and data tools such as Kafka, Spark or Redis.
• AWS or Azure cloud deployment.
• Time series or behavioural anomaly detection experience.
• Exposure to industrial cameras, lighting setup or manufacturing quality systems.

What you get

• Work on enterprise AI projects with manufacturing and large corporate clients.
• Direct exposure to end-to-end delivery, from site visit to go-live.