Cognitive
Architecture
Engineering

Transforming human knowledge into autonomous systems since 2020

Scaled startups to 7-figure ARR through AI automation. Won $4M contracts with first-of-its-kind platforms. Research collaborations with Yale, Microsoft Research, and AutoGPT. 18+ years building production systems.

Pattern Recognition

My LLM obsession began with GPT-3's 2020 release—recognizing that cognition had become programmable substrate. Years before mainstream adoption, I was building production systems, shipping the first generative QA applications, and pioneering RAG architectures.

At UserCue, I engineered AI pipelines that scaled them from $0 to 7-figure ARR, reducing qualitative analysis from 2 weeks to 1 day. At SimSpace, my cyber attack framework won a $4M government contract and now trains defenders globally.

Research collaborations include NeurIPS 2023 with Microsoft Research on AI agent safety, Yale's DAMMA project for Holocaust memorialization, and USENIX publications on ARM security.

I combine deep technical expertise with business impact—18+ years engineering, $11M+ revenue influenced.

18+ Years Engineering
$11M+ Revenue Influenced
95% Time Saved via AI

Engagement Models

AI Feasibility Assessment

$15K

1-2 weeks • Validate before you build

Rapid technical validation with ROI modeling. Know exactly what's possible, what's not, and what it will cost before committing to development.

  • Technical feasibility analysis
  • Data architecture assessment
  • ROI projections with timeline
  • Risk identification & mitigation
  • Full refund if not viable

MVP Development

$50K

4-8 weeks • Proof of concept to pilot

Working prototype that proves business value. From technical validation to production-ready system your team can iterate on.

  • Production-ready MVP deployment
  • Complete documentation & handoff
  • Team training included
  • 30-day post-launch support
  • Fixed price with scope protection

Full System Architecture

$150K+

3-4 months • Zero to production

Enterprise-grade AI platform designed for scale. Complete system with monitoring, observability, and team autonomy from day one.

  • Scalable production architecture
  • Full monitoring & observability
  • Complete knowledge transfer
  • Team becomes fully autonomous
  • Performance guarantees included

Proven With

Startup Founders

Zero-to-one AI products, rapid prototyping, technical risk validation.

Enterprise Leaders

Operational transformation, pilot-to-scale programs, ROI validation.

Research Teams

Theory-to-production bridges, grant technical sections, publication collaboration.

What I Don't Do

No Chatbot Demos

Impressive demos that can't handle real user queries or scale past 10 conversations. I build production systems, not theater.

No Strategy Without Code

PowerPoint 'AI transformation' that never ships. Every recommendation comes with working prototypes and implementation roadmaps.

No Ongoing Dependencies

Maintenance contracts that create vendor lock-in. I document everything, train your team, and make myself obsolete.

Selected Outputs

No-Code AI Pipeline Builder with Custom DSL

UserCue / 4 weeks

Built visual programming pipeline-builder powered by custom DSL compiler that lets non-technical researchers create complex AI workflows. Democratized advanced AI capabilities by making them accessible without coding, enabling domain experts to build sophisticated analysis pipelines.

AI Pipeline Builder Custom DSL No-Code Platform Visual Programming Compiler Design

Technical Innovation Impact

  • Designed and implemented custom domain-specific language for research workflows
  • Designed visual programming interface that compiles to executable AI pipelines
  • Built on top of OpenAI and Anthropic's LLM APIs using bespoke LLM framework
  • Integrated with UserCue's existing data pipeline and data validation system

Tech: Custom DSL, Compiler Design, TypeScript, Vue, AST Processing, OpenAI, Anthropic

Automated Interview Quality Control System

UserCue / 2 weeks

Built AI system that automated quality control processes for qualitative research data, reducing QC time from days to hours while maintaining accuracy at scale. Ensured data quality and consistency across thousands of participants without manual review bottlenecks.

Quality Assurance Process Automation AI Validation Operational Efficiency

Operational AI Impact

  • Reduced quality control time from days to hours through deterministic and non-deterministic validation
  • Maintained data accuracy and consistency across thousands of research participants
  • Minimized manual review bottlenecks in production workflows

Tech: OpenAI, Anthropic, Python, PHP/Laravel, Redis, PostgreSQL

Case study →

Market Research Knowledge Work Automation

UserCue / 4 weeks

Automated complex qualitative analysis workflows that traditionally required expert analysts weeks to complete. Built AI systems that process GBs of research data, extract insights, and generate reports, reducing analysis time by 95%.

95% Time Reduction Knowledge Work Automation Human-in-the-loop Deep Research Agents

Business Impact Through AI

  • Reduced qualitative analysis time from 2 weeks to 1 day (95% time savings)
  • Automated processing of thousands of research participants
  • Built production systems handling complex research workflows at enterprise scale

Tech: OpenAI, Anthropic, Python, Redis, PostgreSQL

Case study →

AI Agent Safety Research

NeurIPS 2023 Workshop / Microsoft Research

Designed and tested safety monitors for autonomous agents in collaboration with Microsoft Research, Northeastern University, and AutoGPT. Defined threat models and demonstrated containment strategies for both inadvertent and intentional malicious activity.

NeurIPS 2023 Microsoft Research AutoGPT AI Safety Threat Modeling

NeurIPS 2023 Publication

Co-designed threat models and safety monitors that intercept unsafe autonomous-agent actions.

Tech: Agent evals, AutoGPT, Python

Read paper →

Available for Select Projects

Working exclusively with teams ready to achieve
10x productivity gains through AI