Senior AI Engineer, Data Gateway

Clearsense; United States; Presencial; USD 140.000 - 190.000; Jornada completa

Descripción de la vacante

Clearsense seeks a senior AI engineer to advance the Data Gateway team by building AI-first retrieval and knowledge-base pipelines on AWS. You will own live project delivery, create reusable tooling, and drive governance, observability, and cost reporting for AI services across data acquisition and curation workflows.

You will design retrieval architectures, map data dictionaries, runbooks, and knowledge bases, and ship agent tools that accelerate engineer productivity while maintaining

Formación

• 5+ years in software, data, or ML engineering with production LLM apps (RAG/Agents).
• Hands-on data work: acquisition, integration, or curation.
• Experience with Amazon Bedrock: model invocation, Knowledge Bases, Agents/AgentCore, Guardrails.
• Built and tuned end-to-end retrieval systems: chunking, embeddings, hybrid search, reranking.
• Strong Python and SQL.

Responsabilidades

• Work on live Data Gateway projects across data acquisition and curation.
• Run first tool implementations on real projects and hand off with playbooks.
• Bring engineering discipline to day-to-day execution: patterns, measurement, AI-first approaches.
• Design and build knowledge bases: source profiles, data dictionaries, runbooks, collections.
• Own the retrieval pipeline end-to-end: parsing, chunking, metadata, embeddings, retrieval.

Conocimientos

LLM applications

Python

SQL

AWS

Retrieval systems

Herramientas

Amazon Bedrock

AgentCore

Python

SQL tools

Descripción del empleo

Department: Product Development, Data Gateway

Reports To: Director, Data Gateway Engineering

About the Role

We are adding AI engineering to the Data Gateway team, the product group that defines how Clearsense acquires, curates, and delivers healthcare data.

This role brings engineering discipline to operations: you will work on live client projects, learn the work firsthand, and reimagine how day-to-day project execution gets done with AI-first approaches. You build the tools, and you run their first implementations on real projects before anyone else does.

You will turn what the Data Gateway team knows (source-system profiles, extraction and curation patterns, data dictionaries, runbooks, and curated document collections) into retrieval-ready knowledge bases, and capture new knowledge as project work happens so every engagement leaves them better.

Then you will build agent tools that make retrieval a routine part of every engineer's day as they discover, map, validate, and document data. Almost all of it runs on AWS, mainly Amazon Bedrock and Amazon Bedrock AgentCore, under our HIPAA and HITRUST controls.

This is a senior individual contributor role. It blends applied AI engineering, hands-on delivery, and a product mindset: do the work, build for reuse, measure quality, and hand off tools other teams can trust.

Success Metrics:

IN 1 MONTH

• Contribute to at least one live Data Gateway project across acquisition and curation, and document where engineers lose time to lookup, mapping, and documentation.
• Inventory the team's knowledge sources and propose a knowledge base design with an evaluation plan.
• Ship a working prototype: one knowledge base and one agent tool used by a Data Gateway Engineer on live work.

IN 3 MONTHS

• Put at least one knowledge base into production with automated ingestion, access controls, and a retrieval-quality baseline scored on every change.
• Deliver at least one agent tool that measurably cuts time on a recurring Data Gateway task, such as source profiling, field mapping, or validation triage.
• Stand up evaluation, observability, and cost reporting for every AI tool the team runs.
• Run the first production use of each tool you ship on a live project, and capture what it learned in the knowledge base.

IN 6 MONTHS

• Maintain a golden path (templates, infrastructure as code, docs) that lets the team stand up a new knowledge base or agent tool in days, not weeks.
• Drive adoption of AI tooling across Data Gateway and at least one Operations Engineering team.
• Keep a living library of retrieval patterns, tool definitions, prompts, and evaluation sets, and pilot at least one emerging AWS capability with an adopt-or-hold recommendation.
• Knowledge capture and retrieval are routine steps in the team's project workflow, not a separate task.

Key Responsibilities:

Hands-On Delivery

• Work on live Data Gateway projects across data acquisition and curation, so you understand the work you are automating.
• Run the first implementation of every tool you build on a real project, then hand it off with a playbook.
• Bring engineering discipline to day-to-day project execution: repeatable patterns, measurement, and AI-first approaches.

Semantic Knowledge Bases

• Design and build the knowledge bases the Data Gateway team relies on: source-system profiles, data dictionaries, extraction and curation patterns, runbooks, and curated document collections.
• Build knowledge capture into project work so source profiles, mappings, decisions, and lessons learned land in the knowledge base as work happens.
• Own the retrieval pipeline end to end: parsing (including OCR'd multi-page PDFs), chunking, metadata, embeddings, hybrid and structured retrieval, and reranking.
• Pick the right store for each corpus, such as Amazon Bedrock Knowledge Bases (managed or custom), Amazon S3 Vectors, Amazon OpenSearch Serverless, Aurora PostgreSQL with pgvector, or graph retrieval on Amazon Neptune Analytics, and document the trade-offs.
• Keep knowledge current with automated sync, versioning, lineage back to source, and retirement of stale content.

Internal AI Agent Tools

• Build agents and tools that help engineers profile sources, map fields, generate and review transformation code, triage validation failures, and draft documentation.
• Build on Amazon Bedrock AgentCore (Runtime, Gateway, Memory, Identity, Policy, Observability, Evaluations), and expose internal systems to agents as MCP tools through AgentCore Gateway.
• Keep a human in the loop where it matters: approval steps, audit trails, and clear limits on what an agent can change.

Quality, Security & Operations

• Define evaluation for every knowledge base and agent: golden datasets, retrieval and answer-quality metrics, and regression checks in CI/CD.
• Apply security by default for PHI: least-privilege IAM, encryption, network isolation, Amazon Bedrock Guardrails, and logging that meets HIPAA and HITRUST evidence needs.
• Track latency, cost, and model usage, and tune model choice, caching, and batch inference to control spend.
• Ship everything as code (AWS CDK or Terraform) through CI/CD.
• Partner with Data Gateway Engineers to find the highest-value problems, and with Operations Engineering to put tools in their hands.
• Make retrieval a routine step in everyone's workflow, built into the tools engineers already use.
• Write documentation, reference architectures, and short training so others can use and extend what you build.
• Work with Security, Compliance, and Infrastructure on reviews and approvals for new AI services and data flows.
• Partner with Product Management to align the AI tooling roadmap with Data Gateway priorities.

Qualifications:

Required

• 5+ years in software, data, or ML engineering, including 2+ years building production LLM applications (RAG, agents, or both).
• Hands-on experience delivering data work (acquisition, integration, or curation), with a clear grasp of how those stages differ and connect.
• Hands-on experience with Amazon Bedrock: model invocation, Knowledge Bases, Agents or AgentCore, and Guardrails.
• Built and tuned retrieval systems end to end: chunking, embeddings, vector and hybrid search, reranking, and evaluation.
• Strong Python and solid SQL.
• AWS depth beyond AI services: IAM, S3, Lambda, Step Functions, ECS/Fargate, VPC networking, and CloudWatch, plus infrastructure as code (CDK or Terraform) and CI/CD.
• Evaluation discipline: you measure retrieval and answer quality and ship against the numbers.
• Experience handling sensitive or regulated data with the right controls.
• You do the operational work yourself before you automate it.
• Clear communicator who documents well and can teach other engineers.

Preferred

• Healthcare data experience: HL7, FHIR, or CCD, and EHRs such as Epic, Oracle Health (Cerner), or MEDITECH.
• Elasticsearch experience, especially semantic or hybrid search over large document indices.
• Document AI: OCR pipelines, Amazon Textract, and layout-aware parsing of scanned clinical PDFs.
• Knowledge graphs or GraphRAG (Amazon Neptune or similar).
• Building MCP servers and tools, and agent frameworks such as Strands Agents or LangGraph.
• Work in HIPAA or HITRUST environments.
• Data engineering background (Airflow, Spark, NiFi) to work alongside Data Gateway Engineers on pipelines.

Why Join Us?

You will reinvent how healthcare data operations run at Clearsense: bringing engineering discipline to day-to-day project work and rebuilding it AI-first.

The knowledge bases and agents you build become the Data Gateway team's shared memory and its force multiplier, so every engineer on Data Gateway and Operations Engineering delivers faster and more consistently.

This is a builder's role with real data, real internal users, and room to set the standard for applied AI at Clearsense.