Design and implement AI workflows on top of existing data pipelines (product feedback extraction, customer update generation, onboarding plans, win/loss summaries, CRM enrichment)
Extend and refine the current TypeScript-based agent: tooling, tool schemas, routing logic, error handling, and observability
Improve transcript and account matching across Zoom, Granola, and Salesforce using entity resolution, heuristics, and/or LLM-assisted matching
Integrate new data sources (Slack, Zendesk, Google Drive, Nexus/telemetry, email) into the existing AWS stack
Define and consume pre-aggregated account/opportunity summaries in S3 for fast, reliable query
Optimize Lambda-based data processing jobs for cost, reliability, and performance
Iterate on model strategy: cheap routing (e.g., Claude Haiku) vs. higher-quality response models (e.g., Claude Sonnet/Opus)
Evaluate prompts, tool selection quality, and response accuracy with clear metrics
Collaborate with GTM stakeholders (Sales, SE, CS, Product, Marketing) to define, test, and refine AI-assisted workflows
Partner with infra engineering (Terraform, Kubernetes) to ensure deployment, security, and observability are production-ready
Contribute to future UI/UX (Slack bot flows, simple web UI/dashboards) to expose workflows to end users

Requirements

5+ years in software engineering, including 1-2+ years building LLM-based applications or agents
Strong experience in TypeScript/Node.js (ideal) or Python for backend and data processing
Hands-on experience with LLM tool calling/agents (custom frameworks, LangChain, or equivalent)
Practical RAG expertise: chunking, metadata schemas, retrieval, relevancy, and evaluation
Solid AWS experience: Lambda, S3, IAM, CloudWatch; exposure to Bedrock is a plus
Background in data integration/pipelines: consuming APIs and webhooks (Salesforce, Slack, Zendesk, Zapier, etc.)
Skill in designing JSON schemas, pre-aggregated summaries, and metadata models for query
Familiarity with Salesforce data structures (accounts, opportunities, leads, activities) or willingness to ramp quickly
Comfortable working in a partially “vibe‑coded” codebase and refactoring toward higher quality (testing, structure, observability)
Experience with prompt engineering and evaluation for business workflows (accuracy, reliability, user trust)

Benefits

Applicant Tracking System Keywords

Tip: use these terms in your resume and cover letter to boost ATS matches.

Hard Skills & Tools

TypeScriptNode.jsPythonLLM-based applicationsAWSLambdaS3JSON schemasdata integrationprompt engineering

Soft Skills

collaborationcommunicationproblem-solvingadaptabilityattention to detail