Capabilities

• Custom model development (computer vision, NLP, time‑series)
• Data pipelines with Airflow / Prefect
• Experiment tracking (MLflow / Weights & Biases)
• Containerized inference (TorchServe / FastAPI + Docker)
• LLM integrations (LangChain, OpenAI, Azure OpenAI)

Reference Tech Stack

• Training: TensorFlow 2.x & PyTorch 2.0
• Serving: FastAPI + Docker + Kubernetes
• Tracking: MLflow / W&B
• Inference: TorchServe / RedisAI
• LLM: LangChain + OpenAI / Azure OpenAI
• MLOps: Airflow / Prefect