About Dice
Dice is the leading career destination for tech experts at every stage of their careers. Our client, CSS Tec, is seeking the following. Apply via Dice today!
Role Overview
You will join a globally distributed engineering organization of fewer than forty people as the tech lead for a scrum team and the people manager for that team. You’ll be accountable for architecture and execution, writing production code, and partnering across Product, Security, and Platform/Reliability to ensure AI features are trustworthy in the real world (evaluation/monitoring, tenant isolation, permissioned tool use, cost controls, and incident-ready operations).
This is a hands-on technical leadership role. Expect to spend significant time designing and shipping production code (Python on AWS). If you are primarily a people manager and are not currently hands-on, this role will not be a fit.
What You’ll Do
Build & Run the AI Platform Layer (Hands-On)
- Lead architecture and delivery of our AI platform services in Python 3.12+ using proven service patterns & platforms (FastAPI, Uvicorn, Pydantic, SQLModel/SQLAlchemy) and production-grade API behavior.
- Own AWS runtime and deployment patterns for the platform: ECS Fargate (API + MCP services), Lambda (doc processing + knowledge ingestion), and event-driven integration via S3 Events and EventBridge.
- Establish “paved road” standards so teams ship safely: service templates, PR/review discipline, CI/CD and environment promotion using Terraform + GitHub Actions (OIDC to AWS), and Docker build practices (including multi-arch where required).
Document Intelligence Pipelines (Warranty Docs → Structured Data)
- Own the end-to-end document processing pipeline (Amazon Textract, Claude Vision).
- Improve extraction quality using deterministic parsing/normalization and custom extractors (e.g., VIN, dates, currency, codes), with strong validation, traceability, and clear failure modes.
- Engineer for reliability and reprocessing: idempotency, bounded retries/timeouts, durable error handling, and controlled replay of failed/changed documents.
Retrieval & Knowledge Base Engineering (RAG That’s Measurable)
- Own the full lifecycle of Amazon Bedrock Knowledge Bases (multiple KBs such as policies/TSBs/procedures/codes): ingestion strategy, change control, and safe promotion across environments.
- Build evaluation and regression testing for retrieval quality (golden sets, automated checks, drift detection) and enforce quality gates so KB changes don’t silently degrade outcomes.
- Implement cost-aware, AWS-native retrieval using Bedrock RetrieveAndGenerate with vector storage in S3 Vectors, and track unit economics (latency and cost per workflow/document/claim).
Agent Workflows & Guardrails (Production, Not Demos)
- Deliver agent orchestration on Bedrock AgentCore Runtime using LangGraph state machines (checkpointing, interrupts, human-in-the-loop steps) with predictable behavior and well-defined failure handling.
- Integrate tools/connectors via the MCP SDK (e.g., DMS connector, VIN decoder, OEM portal tools) with permissioned access, auditable tool calls, and strict boundaries.
- Standardize operational guardrails: CloudWatch logs/metrics (structured JSON logging), security-by-default (Cognito OIDC/PKCE, WAF, Secrets Manager, least-privilege IAM), and runtime discipline for ARM64 AgentCore containers (repeatable builds, including QEMU multi-arch in CI when needed).
The Profile We’re Looking For
You bring strong technical depth, pragmatic leadership, and you can mentor and develop talent. You are proficient being hands-on, and you can scale standards, systems, and a team over time — balancing rapid iteration with production discipline and cost awareness.
- 8–12+ years building and operating backend/platform systems in B2B SaaS.
- Proven hands-on technical leadership for a small team (4–6); comfortable being accountable for architecture and delivery.
- Mastery of Python in production: FastAPI/services, async patterns, workflow orchestration, test discipline, and observability.
- Strong AWS depth (Lambda, ECS/Fargate, S3, IAM, RDS/Postgres, CloudWatch/EventBridge) plus IaC (Terraform preferred).
- Direct experience shipping AI-enabled systems on AWS (Bedrock/RAG, document intelligence such as Textract or multimodal extraction, and evaluation/quality monitoring).
- Experience building production agent workflows with guardrails (permissions, auditability, cost controls, failure modes).
Education
Bachelor’s degree in Computer Science, Engineering, or a related field — or equivalent professional experience delivering and operating enterprise-grade software platforms. Advanced degree is a plus but not required.
