I build and ship production LLM applications—multi-agent systems, RAG pipelines, evaluation frameworks, and real-time inference—running on local GPU infrastructure with zero cloud dependency. I own quality end-to-end from prompt architecture and context design through trust boundaries and reliability checks. I design evaluations against real production scenarios, not synthetic benchmarks, and have enterprise systems architecture experience. My work spans multi-agent coordination, semantic search, multimodal pipelines (audio/text/image/video/music), workflow-driven AI assessment with human-in-the-loop gates, and local inference optimization under real GPU contention.

Mark Lee

PRO

I build and ship production LLM applications—multi-agent systems, RAG pipelines, evaluation frameworks, and real-time inference—running on local GPU infrastructure with zero cloud dependency. I own quality end-to-end from prompt architecture and context design through trust boundaries and reliability checks. I design evaluations against real production scenarios, not synthetic benchmarks, and have enterprise systems architecture experience. My work spans multi-agent coordination, semantic search, multimodal pipelines (audio/text/image/video/music), workflow-driven AI assessment with human-in-the-loop gates, and local inference optimization under real GPU contention.

Available to hire

I build and ship production LLM applications—multi-agent systems, RAG pipelines, evaluation frameworks, and real-time inference—running on local GPU infrastructure with zero cloud dependency. I own quality end-to-end from prompt architecture and context design through trust boundaries and reliability checks.

I design evaluations against real production scenarios, not synthetic benchmarks, and have enterprise systems architecture experience. My work spans multi-agent coordination, semantic search, multimodal pipelines (audio/text/image/video/music), workflow-driven AI assessment with human-in-the-loop gates, and local inference optimization under real GPU contention.

See more

Work Experience

AI Engineer at Driver Recruitment
June 1, 2024 - Present
Designed, built, and shipped a production multi-agent AI platform for a national driver recruitment operation. Processed hundreds of phone calls daily on local GPU infrastructure using a real-time audio pipeline (FreeSWITCH → Whisper STT → local LLM topic classification → observation extraction → SQL persistence). Coordinated specialized agents via a messaging/event-driven architecture with defined capability boundaries. Built a multi-tenant MCP server exposing 40+ tools for semantic search, entity profiles, and workflow management. Implemented a workflow engine using human-in-the-loop draft-then-review patterns and section-based RAG to prevent token overflow while maintaining entity profile fidelity. Delivered production semantic search using RediSearch KNN over candidate/client/job/communications entities, and optimized local LLM inference by testing real-world GPU contention across multiple model sizes. Added LoRA domain distillation and a graduated trust model (AI can read
Senior Architect & Technical Lead at AInrgy
June 1, 2022 - June 30, 2024
Designed and built a scalable IoT telemetry platform for real-time monitoring of distributed solar energy systems. Developed a Node.js microservice backend and managed Raspberry Pi edge devices with CI/CD for field deployment. Implemented automated device discovery and integrated industrial communication protocols (Modbus, RS232). Stored telemetry in InfluxDB and focused on operating under compute/connectivity constraints—architectural patterns later applied to local LLM inference.
Principal Architect & Lead Developer at Doculife Global Inc.
July 1, 2016 - June 30, 2022
Led architecture and development of an enterprise document management platform. Built a Node.js microservice framework with automated service generation from Swagger specs, NGINX orchestration, and database abstraction spanning RethinkDB, Cassandra, and SQL Server. Managed a development team and led CI/CD via Azure DevOps. Decomposed microservices and abstracted protocols, maintained business logic across multiple databases, and integrated encryption for secure document management.
Senior Developer & Architect at Various Companies
January 1, 1991 - June 30, 2016
Provided 25 years of full-stack development and systems architecture across multiple industries and roles. Built and led billing and workflow systems, delivered integrations (including PABX-to-screen for call intelligence foundations), optimized SQL Server performance, and developed a workflow engine. Technologies evolved from VB3/Borland C++ to .NET/C#/Node.js/TypeScript/Python while leading technical direction and team efforts.

Education

Add your educational history here.

Qualifications

Add your qualifications or awards here.

Industry Experience

Software & Internet, Computers & Electronics, Energy & Utilities, Telecommunications, Professional Services, Media & Entertainment
    Autonomous AI Support Agent — Safe by Construction

    An event-driven AI assistant that fields real user questions unattended, in real time — and is read-only by construction: it physically cannot write, commit, deploy, or change data, because those tools simply don’t exist in its toolset. A destructive mistake is impossible, not merely forbidden. It answers only what it has been verified it can answer, cites the source behind each answer, and escalates to a human rather than guessing when something is outside its remit.

    • Real-time, event-driven — replies within human conversational latency
    • Sandboxed by construction — no write path exists, so it is safe to leave running unattended
    • Cites its source; escalates rather than fabricating — no confident-wrong answers
    • Runs on local inference — no cloud, no data egress
    • Stack: TypeScript, tool-calling LLM, event-driven runtime

    An autonomous agent you can actually leave running — because the worst it can do is ask a human.

    Stateful AI for Business-Process Intelligence

    A stateful, agentic system that turns unstructured business interactions — phone calls and SMS — into structured, queryable process intelligence. Unlike tools that treat each interaction in isolation, it preserves semantic context and memory across the whole process, so it understands the story, not just the sentence.

    • Observation extraction: local LLM agents pull structured observations from raw call and SMS transcripts
    • Multi-layered memory (short / medium / long-term) — continuity across interactions, the way a human analyst carries context
    • Ask in plain English: a natural-language query pipeline with progressive refinement over vector embeddings — no query language required
    • All local inference (Ollama) — confidential customer conversations never leave controlled infrastructure; no cloud, no per-token cost
    • Full-stack product: Node.js backend, Angular front end, RESTful APIs, RBAC + JWT, admin and user management
    • Extensible beyond calls to any business process that benefits from persistent, contextual understanding

    Business-process intelligence that remembers — semantic context, not just transcription.

    Full-Stack AI Platform — End to End on Local Inference

    A production platform built end to end for a regulated, compliance-heavy domain — from the database to the UI, with AI woven through, and every model running on local inference so sensitive data never leaves controlled infrastructure.

    • Front end: Next.js (App Router), role-aware single app, shadcn/ui + Tailwind
    • Backend: four TypeScript services (gateway, core, integrations, vector) over tRPC — end-to-end type safety from the SQL query to the React component
    • Data: Postgres + pgvector for semantic search; a fully typed query layer (Kysely + Zod); an audit event on every state change
    • AI, all local: classification, structured extraction, document OCR, agentic tool-calling, and semantic search — zero cloud, no per-token cost, no data egress
    • Audit-grade output: generates regulatory reports validated against the authoritative spec, never fabricating a value it can’t ground
    • End-to-end tested (Vitest + Playwright); pnpm monorepo

    One engineer, the whole stack — a real product where the AI is load-bearing, not bolted on, and the sensitive data never leaves the building.

    Grounded AI Assessment Engine — Cites Its Evidence, Refuses to Guess

    An AI system that evaluates documents and submissions against an authoritative ruleset and returns a structured, defensible determination — with a citation to the specific rule behind every judgement, and a hard refusal to assert anything it can’t ground in the source. Built for high-stakes domains — regulatory, tax, compliance, screening — where a confident wrong answer is a liability.

    • Ingests uploaded documents and structured inputs
    • Assesses against a codified source-of-truth (legislation, policy, rubric)
    • Cites the rule behind each decision; declines when evidence is thin — no hallucinated approvals
    • Full-stack web app; runs on-prem or in the cloud

    Assessments people can trust and audit, because the system shows its work and knows what it doesn’t know.

    Real-Time Voice AI Phone Receptionist

    A production voice AI that answers phone calls in real time — natural conversation, answers service questions, and books appointments end to end, with no human handoff. Built on self-hosted inference with zero third-party AI APIs, so it runs at near-zero cost per call and keeps every second of call data on controlled hardware.

    • Live speech-to-speech over a real phone line (SIP/PBX)
    • Understands intent, handles Q&A, and completes bookings autonomously
    • Zero cloud dependency — self-hosted models; data never leaves the box
    • Stack: Python, telephony/SIP, local LLM inference, real-time audio

    A callable receptionist that never sleeps — at a fraction of a hosted solution’s cost.