As the Azure AI Tech Lead, you will be the principal technical authority for designing, developing, and deploying sophisticated AI and machine learning solutions on the Azure platform.

You will provide technical leadership and architectural guidance to a team of AI/ML engineers, ensuring the successful delivery of complex projects from proof-of-concept to production.

This role requires a deep, hands-on expertise across the AI landscape, from foundational machine learning concepts to the latest advancements in generative AI and MLOps.

You will be instrumental in shaping our technical strategy and fostering a culture of innovation and excellence.

Key Responsibilities :

- Technical Leadership & Architecture: Define and implement robust, scalable, and secure architectures for AI solutions on Microsoft Azure, utilizing services like Azure AI/ML, Azure OpenAI, Azure AI Search, Azure AI Foundry, Azure Cognitive Services and Overall Azure Ecosystem.

- Solution Development & Delivery: Lead the end-to-end development of AI/ML projects, from initial design and data processing to model training, deployment, and ongoing monitoring. This includes hands-on development and guiding the team through complex implementation challenges.

- Team Mentorship: Mentor and guide junior and mid-level AI engineers, fostering their technical growth through code reviews, pair programming, and knowledge sharing sessions.

- DevOps, MLOps & LLMOps: Champion and implement MLOps best practices to establish and streamline CI/CD pipelines for model training, deployment, and monitoring, ensuring reliability and efficiency.

- Innovation & Research: Stay current with the latest advancements in AI, including Large Language Models (LLMs), generative models, and agentic architectures, and drive the strategic adoption of new technologies.

- Cross-Functional Collaboration: Work closely with data scientists, data engineers, product managers, and business stakeholders to translate business requirements into technical solutions and deliver high-impact AI capabilities.

- Governance & Best Practices: Establish and enforce best practices for AI development, including model governance, responsible AI, and performance optimization.

Required Technical Expertise :

Languages & Frameworks :

- Python: Advanced proficiency, including a deep understanding of core ML/AI libraries, advanced data structures, and asynchronous & multiprocessing programming.

- Deep Learning Frameworks: Hands-on experience with PyTorch, TensorFlow, and JAX.

- Ecosystems: Extensive experience with the HuggingFace Transformers & Diffusers ecosystem.

- LLM/Agentic Frameworks: Proven experience with LangChain, LangGraph, LlamaIndex, and Semantic Kernel.

- MLOps: Practical application of MLflow, Weights & Biases, and Kubeflow.

Generative & Agentic AI :

- Retrieval-Augmented Generation (RAG): Deep understanding and implementation experience with standard, graph-based, and vector DB-integrated RAG (FAISS, Pinecone, Weaviate, Milvus).

- Multi-Agent Orchestration: Experience building systems with LangGraph, AutoGen, and tool-calling using OpenAI function APIs.

- Generative Model Fine-Tuning: Hands-on experience fine-tuning diffusion and generative models (e.g., Stable Diffusion, Flux.1, ControlNet, DreamBooth).

- LLM Fine-Tuning: Expertise in parameter-efficient fine-tuning (PEFT) methods such as LoRA and QLoRA on models like Llama3, Mistral, CodeLlama, and Azure OpenAI models.

- Prompt Engineering: Advanced skills in system prompting, function calling, and ensuring safety alignment.

Machine Learning & Deep Learning :

- Classical ML: Strong foundation in tree ensembles (XGBoost, LightGBM, CatBoost), regression models, clustering (k-means, DBSCAN), and dimensionality reduction (PCA, t-SNE, UMAP).

- Deep Learning: In-depth knowledge of CNNs (ResNet, EfficientNet), RNNs/LSTMs, GRUs, and Transformer architectures (BERT, ViT, GPT).

- Graph ML: Experience with Graph Neural Networks (GNNs) using frameworks like PyTorch Geometric or DGL.

- Time Series: Proficiency in forecasting techniques and models (Prophet, ARIMA, DeepAR, Temporal Fusion Transformer).

- Reinforcement Learning: Familiarity with concepts like RLHF, PPO, DQN, and policy gradients.

Computer Vision :

- Object Detection: Experience with models such as YOLO (v5-v8), DETR, and Faster R-CNN.

- OCR: Practical application of PaddleOCR, Tesseract, EasyOCR, and LayoutLM for document understanding.

- Video Analytics: Knowledge of object tracking (DeepSORT, ByteTrack) and other video processing pipelines.

- Multi-modal ML: Experience with models like CLIP, BLIP, Florence-2, and Segment Anything (SAM).

Optimization & Deployment :

- Model Optimization: Expertise in techniques like TensorRT, ONNX Runtime, quantization (INT8, FP16), pruning, and distillation.

- Scalable Deployment: Proven experience deploying models as Dockerized microservices on Kubernetes using REST/gRPC APIs.

- Cloud Platforms: Deep expertise in Azure AI/ML services (AKS, Azure Functions, App Service, Azure OpenAI, AML) and familiarity with AWS SageMaker and GCP Vertex AI and Dataiku.

- Edge AI: Familiarity with deployment on edge devices like Nvidia Jetson or Coral TPU using TensorFlow Lite or CoreML.

- Monitoring: Experience implementing monitoring solutions for model drift and system health using tools like Prometheus, Grafana, and the ELK stack.

Qualifications :

- Education: Master's or Bachelor's degree in Computer Science or a related field.

- Experience: 6+ years of professional experience in AI/ML, with a proven track record of delivering end-to-end solutions.

- Production Experience: Hands-on experience taking ML/LLM solutions from the proof-of-concept stage to scalable production deployment, including monitoring and maintenance.

- Theoretical Foundation: Strong understanding of fundamental ML/DL theory, including optimization algorithms, loss functions, and neural network architectures.

- Problem-Solving: Exceptional analytical, problem-solving, and debugging skills.

- Contributions (A Plus): Demonstrated passion for applied research or engineering through publications, open-source contributions, or patents is highly desirable.