Backend & AI Infrastructure Engineer

Bengaluru, On-site, Full-time, 4–6 years, Immediate joiners preferred

About us

Produsa Labs is a Bengaluru company working on AI products.

The role

You'll own the backend, infrastructure and security behind our AI products. You'll design the systems, run them on AWS, keep them secure and reliable, and keep the AI layer fast and cost-efficient.

What you'll do

• Design and build scalable APIs (REST and GraphQL), services, data models and background job queues.
• Build and run infrastructure on AWS (EC2, ECS/EKS, Lambda, RDS, S3, VPC, IAM), managed as code with Terraform or CDK.
• Own CI/CD: automated tests, safe deploys, quick rollbacks, and separate staging and production.
• Set up monitoring, logging and alerts so problems are caught before users notice.
• Lock down security: OAuth/JWT auth, access control, secrets, encryption, network isolation and rate limiting.
• Run untrusted code safely in isolated containers or microVMs, with strict resource and network limits.
• Integrate LLMs (GPT, Claude and open models), embeddings and vector databases. Route across providers with fallback, caching and cost tracking per request.
• Build agent systems using sound design patterns for autonomous AI behaviour.Qualifications required
• Education: Bachelor's degree in Computer Science, Engineering or a related field.
• Experience: 4–6 years in backend development with Node.js and JavaScript/TypeScript. Python or Go is a plus.
• APIs: Expert at building and consuming RESTful and GraphQL APIs.
• Databases: Strong with relational and NoSQL databases (PostgreSQL, MongoDB).
• Cloud: Hands-on AWS. GCP or Azure is a plus.
• AI: Hands-on with LLMs, embedding models, vector databases (Pinecone, Weaviate, FAISS) and prompt engineering.
• Agents: Solid grasp of agent frameworks, tools and design patterns.
• Security: API authentication and authorization (OAuth, JWT) and security best practices.Preferred
• LangChain, LangGraph, Semantic Kernel or similar orchestration frameworks.
• React.js on the frontend.
• Event-driven architecture and message queues (Kafka, RabbitMQ).
• Docker and Kubernetes. Sandboxing with Firecracker or gVisor.
• MLOps workflows and model lifecycle management.
• Observability tools (Datadog, Prometheus, OpenTelemetry).To apply: apply via LinkedIn and DM us.