AI Engineer

Role: AI ENGINEER
Company: Pushpak AI
Location: India
Employment Type: Full-time
Experience: 1 to 4 Years
ABOUT PUSHPAK AI
Pushpak AI builds AI-powered solutions focused on computer vision, video analytics, video understanding, safety, security, and operational intelligence.
Our solutions are designed for real-world environments across manufacturing, industrial facilities, infrastructure, retail, and enterprise applications. We work on AI systems that process live camera feeds and can be deployed across edge and cloud environments.
ROLE OVERVIEW
We are looking for an AI Engineer who can develop, train, optimize, and deploy computer vision and deep learning models for real-world applications.
The role involves working across the complete AI lifecycle, from understanding a use case and preparing datasets to model development, optimization, deployment, and performance improvement on live camera feeds.
KEY RESPONSIBILITIES
• Develop and train computer vision models for object detection, classification, segmentation, tracking, pose estimation, and activity recognition.
• Build AI solutions for real-time video analytics and video understanding.
• Work with frameworks such as PyTorch, TensorFlow, YOLO, OpenCV, and Hugging Face.
• Fine-tune existing models and develop custom models based on project requirements.
• Prepare datasets, define annotation strategies, perform data cleaning, augmentation, and quality validation.
• Optimize models for real-time inference on CPU, GPU, and edge devices.
• Work with technologies such as OpenVINO, ONNX, TensorRT, CUDA, and NVIDIA GPUs.
• Develop and optimize multi-camera video processing pipelines.
• Implement object tracking and temporal analysis across video frames.
• Evaluate model performance using appropriate metrics and continuously improve accuracy and inference speed.
• Integrate AI models with backend applications, APIs, databases, alerting systems, and dashboards.
• Analyze camera feeds and determine technical feasibility for new AI use cases.
• Participate in Proof of Concepts, pilots, and production deployments.
• Debug model performance issues in real-world environments, including lighting changes, occlusion, camera angles, motion blur, and environmental variations.
• Research and experiment with newer approaches, including Vision Transformers, Vision-Language Models, multimodal AI, and LLM-based video understanding.
REQUIRED SKILLS
• Strong understanding of Machine Learning and Deep Learning fundamentals.
• Strong programming skills in Python.
• Hands-on experience with PyTorch or TensorFlow.
• Good understanding of computer vision concepts and OpenCV.
• Experience with object detection models such as YOLO, Faster R-CNN, DETR, or similar architectures.
• Understanding of object tracking algorithms such as ByteTrack, DeepSORT, BoT-SORT, or similar approaches.
• Experience with dataset preparation, annotation, augmentation, and model evaluation.
• Understanding of model optimization and inference performance.
• Familiarity with Linux and Git.
• Ability to independently investigate and solve real-world AI problems.
GOOD TO HAVE
• Experience with OpenVINO, TensorRT, ONNX Runtime, or CUDA.
• Experience deploying models on NVIDIA GPUs or edge computing systems.
• Knowledge of Vision-Language Models and multimodal models.
• Experience with video streaming protocols such as RTSP.
• Understanding of multiprocessing, multithreading, GPU utilization, and high-throughput inference pipelines.
• Experience with cloud platforms such as AWS, Azure, or GCP.
• Familiarity with Docker and production AI deployment.
• Experience working with industrial CCTV, surveillance systems, or large-scale video analytics platforms.
PREFERRED BACKGROUND
• B.Tech, B.E., M.Tech, M.Sc., or equivalent qualification in Computer Science, Artificial Intelligence, Data Science, Electronics, or a related field.
• 1 to 4 years of relevant experience in AI, Machine Learning, Computer Vision, or Deep Learning.
Candidates with strong practical experience, relevant projects, or demonstrated AI engineering skills are encouraged to apply even if they do not meet the exact experience requirement.
WHAT WE LOOK FOR
We are looking for engineers who take ownership of problems and can translate real-world requirements into reliable AI solutions.
The ideal candidate should be comfortable experimenting with different approaches, evaluating trade-offs, improving model performance, and building solutions that can operate reliably in production environments.
Strong problem-solving ability, attention to detail, and a practical approach to AI development are important for this role.

Work arrangement
No