David Robertson

Lead AI software engineer

I build AI products that need to survive real users, messy domains, and production constraints.

I work across LLM application architecture, agent systems, RAG workflows, product strategy, and hands-on delivery. I am actively looking for my next senior or lead role and can start fast.

David Robertson
10+
years shipping production software
4+
years focused on LLM systems
50%
faster dev iteration on a LangGraph RAG pipeline
<400ms
launch-day P95 on GKE agent infrastructure

Recent work

CogniformAI

Lead AI Software Engineer

2024 to present

Leading product strategy, architecture, and delivery for an LLM-powered SaaS form platform built with Nuxt, FastAPI, Postgres, and GCP.

Waymark

AI Engineer

2025 to present

Designing an AI-accelerated commercial creation pipeline spanning multiple LLMs, generative video, timeline assembly, and natural-language project editing.

Vantager

Generative AI Engineer

2025

Shipped a LangGraph microservice for private equity workflows that cut investor report generation from 5+ minutes to under 20 seconds.

FormKit

AI Integration Specialist and Lead Developer

2023 to 2024

Conceived and shipped an AI form builder end-to-end, trained the engineering team, and helped turn the work into a flagship AI product line.

Proof

What I am best at

Turning fuzzy AI ambition into a shipped product. That usually means clarifying product boundaries, designing systems that remain debuggable after launch, and moving between UI, APIs, orchestration, and infrastructure without creating handoff drag.

Selected outcomes

  • Reduced investor report generation time by 93 percent in a private equity workflow.
  • Cut form-creation operations by 70 percent through AI-assisted generation.
  • Published a Modular guest post after beating Unsloth's CUDA kernel with zero GPU experience.

Core stack

Nuxt, TypeScript, Python, FastAPI, Postgres, GCP, LangGraph, RAG pipelines, OpenAI, Anthropic, and production AI workflow design.

Contact

If you are hiring for senior or lead AI engineering, I would love to talk.

Fastest path is email. I am also happy to connect through LinkedIn, GitHub, or X.