Lead Software Engineer at SalesAi · Chicago

Dan Lewis

I build production AI agents and the platforms underneath them.

LLM-driven SMS and voice agents, workflow engines and public APIs, plus about 13 years of full-stack work across payments, finance and advertising.

Selected work

AI SMS agent

Compliance before the model

Designed, built and owned an AI SMS agent end to end. TCPA consent and do-not-call checks run before any LLM call, and detect-and-regenerate guards catch hallucinated replies.

Voice-agent memory

Context without per-turn latency

Architected cross-call contact memory with tenant-scoped pgvector retrieval. Context is assembled once at call start, trading some mid-call freshness for zero added latency on each turn.

Voice platform runtime

Guardrails that fail safe

Engineered the workflow engine and voice-agent runtime for a multi-tenant platform. Removed misdispatched calls and concurrency leaks, and made guardrail failures fail safe instead of stalling a live call.

Public API

Opening agents to customers

Led the design and build of a public API for AI agents and workflows, with gateway routing, token exchange, per-credential rate limits, idempotency and revocation.

Experience

2026 – present

Lead Software Engineer · SalesAi

  • Own the SMS agent, the public API, and workflow-engine and voice-agent correctness.
  • Evaluate agent quality with eval sets, regression runs on prompt changes and transcript review.
  • Set technical direction across teams through design reviews and ADRs, and mentor engineers at varying seniority.
2025 – 2026

Senior Software Engineer · Meta

  • Developed REST APIs and data pipelines that let ad agencies publish creator-driven ads at scale, driving $162M in ARR.
  • Designed and built an ML-powered brand classification tool for multi-brand retailers, addressing $300M in revenue and performance gains.
  • Created and led a program to find missed integration opportunities with mid-sized clients, adding $28M in ARR.
2023 – 2025

Senior Software Engineer · Golub Capital

  • Created APIs for data queryability, entitlements and work-item tracking, so non-technical staff could self-serve requests that used to need a ticket.
  • Planned and executed .NET application and database migrations to Azure, cutting hardware costs and improving uptime.
2018 – 2023

Software Engineer · Discover Financial Services

  • Developed APIs connecting ServiceNow with Hazelcast for automated failover and failback, protecting Discover's payment processing system from disruption.
  • Provided SME support on production payment and settlement issues for critical applications.
2013 – 2018

Earlier

  • Freelance: designed and ran theatrical control systems for universities, architecture firms and museums, including a TwinCAT GUI that flew a live performer.
  • Vision2 Systems: built a client-facing full-stack payment processing app for nonprofits and churches, and added Elasticsearch to cut search times.

Skills

AI & LLM

  • LLM applications
  • Conversational & voice agents
  • Tool calling / MCP
  • Guardrails
  • RAG & pgvector
  • Prompt engineering
  • Evals
  • LiveKit
  • OpenAI
  • Claude
  • Gemini

Languages & frameworks

  • Java 21
  • Spring Boot
  • C# / .NET
  • TypeScript
  • Node.js
  • React
  • Next.js
  • Python
  • SQL

Cloud & infrastructure

  • Google Cloud
  • Azure
  • Terraform
  • Docker
  • Kubernetes
  • PostgreSQL
  • Redis
  • Elasticsearch
  • Twilio
  • Auth0
  • Datadog

Leadership

  • Design reviews & ADRs
  • Mentoring
  • Cross-team technical direction
  • Shape Up
  • Client integrations

Education

University of Wisconsin–Madison · BS, Design & Theatre