AI Architect

Location: Ahmedabad, India

Department: Product Development

Senior AI Systems Architect (On-Premise & Secure Environments)

Role Overview

Key Responsibilities

• System Design: Architect the end-to-end lifecycle of AI products, from data ingestion and embedding generation to front-end delivery.
• Secure Deployment: Lead the transition of AI models from development to on-premise, air-gapped environments, ensuring zero external dependencies.
• Infrastructure Orchestration: Design and manage containerized microservices using Docker and Kubernetes (K8s) optimized for local hardware.
• Full-Stack Integration: Collaborate with engineering teams to integrate Python-based AI services with React-based frontends via high-performance APIs.
• Data Strategy: Implement and optimize Vector Databases and traditional relational databases to support RAG (Retrieval-Augmented Generation) workflows.
• Security First: Implement rigorous security protocols, including encryption at rest/transit, identity management, and model weight protection within restricted networks.

Technical Requirements

AI & Data Science

• Expertise in Python (FastAPI, Flask, or Django) for building scalable AI services.
• Deep understanding of Embeddings, Vector Search (e.g., Milvus, Qdrant, Weaviate), and LLM orchestration.
• Experience fine-tuning or deploying open-source models (Gemma, etc.) locally.

Architecture & DevOps

• Containerization: Mastery of Docker and orchestration via Kubernetes.
• Deployment: Proven track record of On-premise deployments and managing "Sneakernet" or air-gapped software update cycles.
• APIs: Experience designing secure, versioned RESTful or GraphQL APIs.

Frontend & Databases

• React: Ability to architect how frontend applications consume complex AI streaming data.
• DB Management: Proficiency in PostgreSQL, NoSQL, and specialized Vector DBs.

Security & Networking

• Experience with hardened Linux environments.
• Knowledge of network security in restricted environments (firewalls, proxy management, and certificate handling).

Preferred Qualifications

• Experience with GPU acceleration (CUDA/Triton) in local environments.
• Knowledge of MLOps tools adapted for offline use.
• Background in highly regulated industries (Defense, Healthcare, or Finance).