Artificial Intelligence Engineer

Role: Vision Analytics AI Engineer

Location: Technopark, TVM

Role: Lead AI Engineer (6–10 Years) | Senior AI Engineer (3–6 Years)

About the Role

We are seeking a specialized AI Engineer to lead the development of our next-generation Vision Analytics platform. You will be responsible for building intelligent systems that can "see," "understand," and "reason" about visual data. This role requires a unique blend of traditional Computer Vision expertise and cutting-edge Generative AI (Multi-modal LLMs/VLMs) to solve complex industrial and enterprise challenges.

Key Responsibilities

1. Advanced Vision Analytics & Deep Learning

• Architect & Deploy: Design and implement state-of-the-art Computer Vision models for object detection, segmentation, tracking, and activity recognition.
• Video Intelligence: Build real-time video analytics pipelines to extract actionable insights from CCTV and industrial camera feeds.
• Optimization: Hardening and optimizing models for edge deployment (NVIDIA Jetson, OAK-D) and cloud environments.
• Synthetic Data: Use Generative AI/GANs to create synthetic datasets for training vision models in data-scarce environments.

2. Generative AI & Multi-modal Integration

• Vision-Language Models (VLMs): Implement and fine-tune models like CLIP, LLaVA, or GPT-4V for visual Q&A and image-to-text reasoning.
• Multi-modal RAG: Develop Retrieval Augmented Generation systems that can query both text and image/video metadata using vector databases (Milvus, Pinecone, or Weaviate).
• Visual Agents: Contribute to Agentic AI systems that use vision as a primary input to trigger automated workflows or alerts.

Required Skills & Experience

1 Computer Vision Core:

• Expertise in OpenCV, MediaPipe, and SOTA architectures (YOLOv8/v10, Detectron2, SAM - Segment Anything Model).
• Strong proficiency in python and ML Frameworks PyTorch , TensorFlow/Keras, OpenCV .

2 Generative AI & Multi-modality:

• Hands-on experience with Large Language Models (LLMs) and Vision Transformers (ViT) .
• Experience with RAG frameworks (LangChain, LlamaIndex) specifically applied to multi-modal data.
• Proficiency in Prompt Engineering for both text and vision models.
• Proven experience building Agentic Workflows using Popular Agentic Frameworks like LangGraph, AutoGen, CrewAI, or PydanticAI.

3 Programming:

• Strong Python skills (NumPy, Pandas, Scikit-Learn).
• Familiarity with model quantization (TensorRT, ONNX, OpenVINO).

Good to Have

• Experience with MLOps (MLflow, DVC) and CI/CD pipelines for Vision models.
• Knowledge of NVIDIA DeepStream or GStreamer for video pipeline orchestration.
• Cloud AI experience (Azure Vision, AWS Rekognition, or GCP Vertex AI).

Educational Qualification

• Bachelor’s or Master’s degree in Computer Science, AI, Robotics, or a related field.