AI Engineer III – Agentic Platform

Overview

Build and scale the headless, MCP-native Agentic Platform behind Amtech AI Intelligence. Set technical direction, make build-versus-buy decisions on agent tooling, and provide technical guidance to Forward Deployed Engineers embedding the platform into product pods.

What you'll do

• Build business logic that runs independently of any screen for use by people in the UI, AI agents, and automated jobs.
• Expose Amtech capabilities as MCP-described tools for AI model access.
• Build and maintain the LLM orchestration layer that turns model output into safe, governed, tool-using actions.
• Implement and extend agent workflows using frameworks such as LangChain/LangGraph and the Model Context Protocol (MCP).
• Build and tune RAG pipelines to ground agent responses in Amtech data.
• Integrate model-agnostic LLM providers through a common provider abstraction.
• Build and maintain the Data Agent to read EnCore and other customer data directly for inferencing without dependency on legacy point-to-point APIs.
• Ensure every agent action passes governed approval gates and has a complete, queryable audit trail while customer data remains in the customer's environment.
• Extend the platform into Schedule Intelligence, Order & Customer Intelligence, Docs & Knowledge, and Forecast & KPI Intelligence capabilities.
• Support the Utilization mandate by enabling agent development for Sales, Support, and Customer Success on the shared governed orchestration, RAG, and audit foundation.
• Turn Amtech's first production agent pattern into a reusable platform component for other product pods.
• Partner with Forward Deployed Engineers to validate the platform against real workflows and feed field learnings into the roadmap.
• Deploy agents and inference endpoints on AWS and integrate them into existing APIs and microservices.
• Set up evaluations, cost controls, and basic observability to identify regressions, drift, and runaway spend before they reach customers.
• Partner with Data Science, ML Engineering, and MLOps on feature stores, model registries, model CI/CD, drift monitoring, deployment, monitoring, and retraining across agentic and classical ML workloads.
• Set technical direction and review the work of other engineers.

What you'll need

• Bachelor's Degree in Computer Science, Software Engineering, Data Science, or a related field.
• 4-6 years of software engineering experience, including 2+ years hands-on building production systems with large language models.
• Demonstrated experience setting technical direction and reviewing the work of other engineers, even without a formal people-management role.
• Strong Python programming skills.
• Practical experience with prompt engineering, tool/function calling, and agent frameworks such as LangChain and LangGraph.
• Experience integrating at least one major LLM provider API: OpenAI, Anthropic, Azure OpenAI, or AWS Bedrock.
• Hands-on AWS experience, such as Lambda, ECS/EKS, S3, IAM, and CloudWatch, deploying and operating production workloads in the cloud.
• Experience with real-time and batch model inferencing, including optimizing latency, throughput, and cost for LLM or ML workloads.
• Experience building or consuming RAG pipelines and working with vector databases.
• Working knowledge of classical ML engineering and MLOps practices, such as feature stores, model registries, and MLflow/Kubeflow-style CI/CD for models, enough to partner with Data Science, ML Engineering, and MLOps, not necessarily with hands-on ownership.
• Familiarity with Docker and containerized deployment.
• Working knowledge of Git and CI/CD practices.

Nice to have

• Experience with the Model Context Protocol (MCP) or similar agent-tool integration standards.
• Experience building governed, read-only data access layers with audit logging, guardrails, and least-privilege access.
• Exposure to Kubernetes and cloud AI/inferencing platforms such as AWS Bedrock and AWS SageMaker.
• AWS certification is a plus, not a requirement.
• Experience working with ERP, MES, or other industrial/manufacturing data environments.
• Prior experience embedding directly with a product team or customer-facing engineering effort.
• Familiarity with performance monitoring tools such as Prometheus, Grafana, and Datadog.

Details

• Location: Bangalore.