SAIC is seeking an Automation & AI Engineer to help modernize legacy systems into secure, cloud-native applications for a federal agency that processes high-value financial transactions. This hybrid role in Washington, DC focuses on building practical LLM and agentic AI capabilities, integrating legacy data, and applying platform engineering plus MLOps/AIOps practices to production environments.
Compensation: USD 120,001 - 160,000 per year.
Responsibilities
- Modernize GMF and related legacy workloads by refactoring monoliths and batch processes into secure, cloud-native architectures using microservices, APIs, and event-driven systems with embedded AI and automation.
- Design, build, and deploy LLM- and agentic AI–based solutions (including LangChain, LangGraph, RAG, vector search, and AWS Bedrock agents) to automate complex workflows and integrate with IRS data sources.
- Implement platform engineering and MLOps/AIOps best practices, including CI/CD, infrastructure-as-code, model/prompt lifecycle management, and responsible AI controls.
- Partner with architects, developers, testers, and stakeholders to deliver scalable and secure AI-driven modernization solutions.
- Integrate legacy data sources into modern data platforms and AI-enabled services.
- Support compliance with security, privacy, and governance needs in a regulated federal financial environment.
Requirements
- Bachelor’s degree in Computer Science, Engineering, Data Science, or related field.
- Ability to obtain and maintain a public trust requiring U.S. Citizenship or Green Card.
- 9+ years in software, ML, or data engineering, including experience with application modernization.
- 4+ years building and deploying AI/ML or LLM-based applications in production.
- Strong experience with modern application architectures (microservices, REST APIs, event-driven) and legacy integration.
- Hands-on experience building agentic AI solutions using LLM frameworks such as LangChain and LangGraph.
- Proficiency in Python and common ML/NLP libraries including Hugging Face, Transformers, scikit-learn, and PyTorch/TensorFlow.
- Production experience with AWS services including networking/IAM, Lambda, ECS/EKS, API Gateway, S3, DynamoDB, RDS, OpenSearch, SageMaker, and CloudWatch.
- Practical experience using AWS Bedrock for LLM applications and agents (including knowledge bases and guardrails).
- Experience implementing RAG and working with vector search (vector databases/search systems).
- Experience with CI/CD and infrastructure-as-code such as Terraform or CloudFormation.
- Familiarity with MLOps/AIOps (for example, MLflow and SageMaker) and AI-focused observability including logging, metrics, and drift/quality monitoring for LLM/agent workflows.
- Strong SQL skills and experience integrating legacy data into modern platforms.
- Experience with Docker and container orchestration (Kubernetes, AWS ECS/EKS).
Technologies
Python, Hugging Face, Transformers, scikit-learn, PyTorch, TensorFlow, AWS, AWS Bedrock, LangChain, LangGraph, RAG, vector search, Terraform, CloudFormation, MLflow, SageMaker, LangSmith, Databricks, CrewAI, AutoGen, Temporal, Model Context Protocol (MCP), BrainTrust, DeepEval, CI/CD, microservices, REST APIs, event-driven, Docker, Kubernetes, ECS/EKS, Lambda, API Gateway, S3, DynamoDB, RDS, OpenSearch, CloudWatch
Desired
- Strong technical judgment and communication skills, including the ability to explain AI modernization approaches to technical and business stakeholders.
- Experience with Databricks (notebooks, Delta Lake, ML/feature store) for data and ML pipelines.
- Experience with LLM/agent observability and debugging tools (for example, LangSmith or similar).
- Experience with advanced agent frameworks such as CrewAI and AutoGen, including multi-agent workflows.
- Hands-on experience operating agents in production, including safety/guardrails, performance tuning, and lifecycle management.
- Experience with durable workflow engines such as Temporal for long-running AI/automation workflows.
- Familiarity with Model Context Protocol (MCP) for tool integration and extensible agent systems.
- Experience with LLM/agent evaluation frameworks such as BrainTrust or DeepEval (or similar).
