2 days ago
USA (city not specified)Senior / Staff+
H1B sponsor
Responsibilities
- Stand up the Amazon Bedrock Knowledge Base and Amazon OpenSearch Serverless vector collection using infrastructure code, embedding models, chunking strategies, and metadata fields.
- Implement hybrid vector and keyword retrieval with taxonomy, source, status, review-date, and entitlement metadata filtering, plus top-k, threshold, and reranking tuning.
- Build a versioned Python retrieval API using API Gateway and Lambda with authentication, rate limits, usage metrics, structured errors, and citation payloads.
- Enforce query-time entitlement filtering so restricted passages are not returned to unauthorized callers.
- Provide a Bedrock RetrieveAndGenerate reference endpoint for pilot validation.
- Build dashboards for relevance, grounding, citation quality, permission tests, freshness, latency, errors, and OpenSearch capacity.
- Support the contact center pilot by capturing retrieval misses, tuning the system, and rerunning evaluation suites.
- Write retrieval and API runbooks covering model changes, re-embedding, index rebuilds, and cost implications.
- Pair with client platform owners and complete the transition readiness handoff.
Requirements
- 4+ years of experience as a software or ML engineer.
- At least one production RAG or enterprise search system delivered and measured with an evaluation suite.
- Hands-on experience with Amazon Bedrock Knowledge Bases, Retrieve, RetrieveAndGenerate, and embedding models.
- Experience with Amazon OpenSearch Serverless vector search, hybrid queries, and metadata filters.
- Python experience shipping Lambda-based APIs behind API Gateway, including authentication, versioning, and throttling.
- Infrastructure-as-code experience with Terraform or AWS CDK and CI deployments into client-owned accounts.
- Practical knowledge of chunking, embedding model selection, reranking, and relevance, latency, and cost trade-offs.
- Experience building operational dashboards in CloudWatch or an equivalent tool.
- Preferred: AWS Certified Generative AI Developer Professional or AWS Certified Machine Learning Engineer Associate.
- Preferred: experience with RAG evaluation frameworks and faithfulness, context precision, context recall, and citation accuracy metrics.
- Preferred: experience integrating with a contact center Agent Assist product as an API provider.
- Preferred: familiarity with Bedrock Guardrails and permission-aware retrieval over Microsoft 365 content.
Benefits
- Opportunity to shape real-world AI-driven projects across industries and work with clients from startup innovation through enterprise transformation.
- Global collaboration across continents and cultures.
- Inclusive environment supporting continuous learning, innovation, and ethical AI standards.
Categories
AI ApplicationsData Engineering
About Solvd
Solvd is a privately held software engineering and consulting firm that builds AI, data, cloud, application development, and quality engineering solutions for enterprises across technology, ecommerce, retail, fintech, hospitality, and banking. Headquartered in Walnut Creek, California, it is a portfolio company of Siguler Guff & Company. Following its acquisition of Tooploox, Solvd delivers end-to-end engagements from strategic advisory to custom AI development and enterprise-scale implementation for Fortune 500 clients.