AI Engineer

Role Description

This is a full-time, on-site AI Engineer role based in Hyderabad. The AI Engineer will own and evolve VelsolVision, a production, GPU-accelerated computer-vision service that ingests fleet camera video, runs real-time deep-learning inference, and detects construction-site safety violations (such as missing hardhats and safety vests) at scale. Day-to-day responsibilities include training, fine-tuning, and evaluating computer-vision models (YOLOv8 / PyTorch), curating and labeling datasets, and improving detection accuracy through temporal-smoothing and confidence-scoring algorithms. The role involves maintaining an event-driven, asynchronous Python microservice that consumes Azure Event Hub events and orchestrates video download, GPU inference, result aggregation, and persistence across gRPC, Cosmos DB, and REST integrations. The AI Engineer will collaborate with product managers, platform engineers, and data teams to translate safety and compliance requirements into scalable technical solutions, operate the service on GPU-enabled Kubernetes, and sustain model performance through evaluation, monitoring, and iterative retraining. The role also includes contributing to automated testing, code reviews, documentation, and AI/ML best practices across the organization.

Qualifications

• 8+ years of experience in software and AI engineering, with a strong foundation in Computer Science, including data structures, algorithms, and software engineering principles.

• Proficiency in Computer Vision and Neural Networks, with hands-on experience designing, training, fine-tuning, and evaluating deep-learning models (object detection such as YOLO/Ultralytics, and frameworks such as PyTorch).

• Practical experience building and curating datasets, labeling, and tuning models for accuracy (precision/recall, confidence and IoU thresholds) on noisy, real-world video.

• Strong Python software development skills, including asynchronous programming (asyncio) and building production-grade, service-oriented systems (e.g., FastAPI).

• Experience with large-scale, event-driven data pipelines and distributed-systems fundamentals — message/event streaming (Azure Event Hub, Kafka, or similar), at-least-once processing, checkpointing, and idempotency.

• Hands-on experience with cloud platforms (Azure preferred; AWS or GCP acceptable) and cloud services such as Cosmos DB, Blob Storage, and Managed Identity.

• Experience deploying and operating containerized workloads on Kubernetes, including GPU-accelerated inference (NVIDIA), Docker, and GitOps tooling (Flux/Argo).

• Familiarity with gRPC/protobuf and REST API integration.

• Bachelor’s or Master’s degree in Computer Science, Artificial Intelligence, Machine Learning, or a related technical field.

• Ability to work collaboratively in cross-functional teams, with strong analytical, problem-solving, and communication skills.

• • Prior experience with video analytics, safety/compliance detection, image processing (OpenCV), or Azure Custom Vision is a plus.