Applied AI / LLM Systems
Building with NLPg | AI-assisted SDLC | Business, Product & AI Enablement
When it comes to AI, the person who can best describe the problem is the person who will solve it.
I believe that English is becoming one of the hottest new programming languages.
I work with AI systems, business intent, product thinking, documentation, and AI Enablement.
My focus is helping teams move from AI experimentation to structured implementation by turning intent, constraints, workflows, and context into serious system inputs.
In practical terms, I work on:
- translating business problems into AI-ready workflows
- shaping requirements, prompts, policies, and documentation into operational material
- designing governed LLM workflows with human decision gates
- building AI-assisted SDLC experiments and prototypes
- making AI systems easier to inspect, explain, evaluate, and adopt
I am less interested in making AI sound impressive than in making it work under conditions that resemble real life.
Deterministic Agentic Execution for AI-assisted SDLC
DAX is my core project.
It explores how AI-assisted work can become more structured, inspectable, and governed by treating AI execution as a controlled workflow rather than an open-ended chat experience.
The problem DAX addresses:
AI tools are powerful, but real delivery environments need intent capture, constraints, approvals, auditability, recovery, and human control.
The mechanism:
DAX uses deterministic execution patterns, RAO loops, approval gates, audit trails, local-first control, and structured workflows to make AI-assisted work more operationally useful.
The larger thesis:
AI-assisted SDLC will not mature through better prompts alone. It needs orchestration, governance, documentation, evaluation, and clear human intervention points.
DAX represents my main professional thesis:
AI should not just generate. It should operate inside systems that can be inspected, corrected, governed, and trusted.
DAX is the center. The surrounding projects support the larger thesis of governed, explainable, and usable AI work.
| Project / Area | Focus | Purpose |
|---|---|---|
| DAX | Deterministic AI-assisted execution | Core project for governed AI workflows and AI-assisted SDLC |
| Context Engineering Core | Intent-first programming and structured prompt logic | Framework for aligning LLM behavior with business goals, constraints, and operational reality |
| Soothsayer / Rook experiments | Agent workflows, CLI/TUI interaction, orchestration patterns | Exploration of how AI systems should communicate, execute, and expose state |
| Local SLM / MLX experiments | Local inference, privacy-aware intelligence, small model workflows | Understanding cost, latency, privacy, and control trade-offs |
| PruningMyPothos | AI writing, documentation, searchability, and thinking systems | Public writing space for systems, sentences, and sentiments |
I am currently focused on how orchestration and control models change the practical usefulness of AI systems.
Areas I actively explore:
- deterministic execution patterns for agentic systems
- AI-assisted SDLC and developer workflow augmentation
- Natural Language Programming as a practical interface for system design
- context engineering, prompt architecture, and guardrail design
- evaluation methods that go beyond “it feels smarter”
- local and privacy-aware AI workflows using small language models
- Apple MLX and local inference trade-offs
- documentation systems that improve AI adoption and searchability
- ways to make AI systems legible to non-developers without flattening technical truth
Research output is always tied back to adoption, risk, delivery, and operational reality.
My work aligns with roles and problems where GenAI, business analysis, product thinking, delivery, documentation, and implementation overlap.
Relevant role families:
- GenAI / LLM Consultant
- AI Business Analyst
- AI Product or Program Consultant
- AI Implementation Consultant
- AI Solutions Consultant
- AI Enablement and Adoption Lead
- AI Workflow and Governance Consultant
- Technical Product Documentation Specialist
- AI-assisted SDLC / LLMOps Analyst
The common thread is simple:
I help teams translate business intent into governed AI workflows, implementation-ready requirements, documentation systems, and working prototypes.
- Why a good PRD is often closer to source code than most teams admit
- The difference between prompting as guesswork and context engineering as disciplined setup
- Building AI systems where humans remain meaningfully in the control loop
- Making orchestration, governance, and traceability feel usable rather than bureaucratic
- Helping non-developers work with serious AI systems without pretending to be software engineers
- Making AI outputs easier to evaluate, explain, and operationalize
I play guitar, love dogs, and spend downtime listening to metal.