Hitachi Digital Services is looking to raise engineering maturity by driving standards, guidelines, platform capabilities, and large-scale technical debt remediation. The job aims to build advanced agentic AI workflows to automatically analyze codebases, detect tech debt, and generate high-quality fixes.
Requirements
- Expertise in Python programming, software engineering best practices, testing strategies, CI/CD, and system design.
- Hands-on experience shipping LLM-powered features such as autonomous workflows or function calling with measurable impact on reliability or latency.
- Deep knowledge of multi-agent architectures including planners, executors, and tool routing.
- Strong understanding of RAG systems: chunking, embeddings, vector/hybrid search, and retrieval policies.
- Experience evaluating LLMs and agent workflows incorporating statistical reasoning and validation.
- Proficiency with AWS (Lambda, ECS/EKS, S3, API Gateway, EC2, IAM) and Infrastructure-as-Code for cloud resource automation and deployment.
- Experience with observability tools (Datadog, logging, tracing, metrics).
Responsibilities
- Design, develop, and maintain LLM-powered multi-agent workflows for code analysis, remediation proposals, and safe patch generation.
- Implement agentic patterns including planning/execution loops, dynamic tool orchestration, sandboxing, guardrails, and failure recovery.
- Build scalable automation systems for technical debt remediation: language/runtime upgrades, vulnerability patching, dependency modernization, and config drift correction.
- Architect and optimize Retrieval-Augmented Generation (RAG) pipelines, managing chunking, embeddings, hybrid search, reranking, and retrieval policies.
- Develop robust evaluation frameworks for LLMs, RAG, and agent workflows, including offline datasets, validation metrics, statistical testing, and A/B tests.
- Contribute to backend systems using Python, distributed systems, microservices, PostgreSQL, DBT, vector databases, caching, streaming, and queueing.
- Build CI/CD pipelines, observability dashboards, and perform performance analysis on model, retrieval, and network layers.
Other
- 5+ years experience building production-grade systems with end-to-end ownership.
- Partner with Dev Experience and Platform teams to define engineering guidelines and reusable standards across the organization.
- Collaborate cross-functionally with product, platform, and security to move prototypes to production-grade services.
- Communicate clearly with stakeholders, write technical documentation, and mentor junior engineers.
- Adaptable and proactive problem solver.
