Strategy · Analytics · Applied AI

I build clarity into complex systems.

Independent product builds — AI in the loop, operator judgment in the seat. Informed by a decade across enterprise systems, from actuarial analytics to billing, pricing, fare logic, and platform migrations.

View Work Get in Touch

Clarity in ambiguity

Turning messy problems into structure.

Decision-grade analytics

Insight that holds up to interrogation.

Auditable AI

Inputs traced. Outputs explainable.

§ 02Selected Work

Selected projects.

Independent builds in strategy, analytics, and applied AI for enterprise and operational contexts.

Case Study / GridPulse

GridPulse

An integrated decision platform for power markets — unifying weather forecasts, scenario analysis, and grid telemetry across operating roles.

An independent build exploring decision-support design for power markets. Weather, forecasts, and scenario analysis consolidated into role-based operating views for traders and grid operators — 51 US balancing authorities (~99% of lower-48 demand), four-model ensemble with real holdout metrics, deployed on Cloud Run with scheduled scoring + training jobs.

51 BAs · ~99% lower-48XGBoost · Prophet · SARIMAX ensembleReal holdout metricsCloud Run + scheduled jobs

Read the case study →Open ↗Source ↗

All Work12 projects

01ValuateAI-augmented DCF agent — extracts financial line items from SEC 10-Ks via XBRL and Claude, then runs a Monte Carlo valuation with cell-level source attribution.AI Financial Agent 02TenancyLease abstraction agent for multifamily operators — extracts nine structured sections from residential lease PDFs with per-field source citations, click-to-highlight overlays anchored to OCR coordinates, a human-review exception queue, and grounded Q&A. The same agent is exposed as an MCP server.AI Document Agent 03Eval HarnessDefensible cost/quality eval comparing 9 LLMs (5 open-weight on DGX Spark via Ollama, 4 closed-weight via API) across 4 real production tasks from Sift. Cross-vendor judging, hardware-amortized cost, verifiable held-out lock. Routing-decision framework, not a benchmark.Applied LLM Evaluation 04Quantization StudyControlled experiment on Llama 3.1 8B at FP16 / Q8_0 / Q4_K_M — paired design across MMLU and CoNLL-2003 NER (n=2,400). Decision framework, not a leaderboard.Applied LLM Research 05TarazuAI-assisted prioritization platform applying the RICE framework, with interactive scoring and structured strategic guidance.Product + AI 06SiftA news aggregator with civic footnotes. Reads ~50 outlets across the political spectrum, AI-summarizes today's stories across 10 categories — and on top, every politician, organization, bill, and political term in an article links to a structured dossier sourced from public records.AI News + Civic Literacy 07FocusForgeCharacter-driven Pomodoro for iOS that translates focus sessions into RPG-style character progression. Template-based on-device AI coach for intent framing, post-session reflection, and streak rescue. Built in SwiftUI; feature-complete and entering beta.iOS Mobile App 08GTM Healthcare IntelligenceSpecialty-EHR go-to-market analytics architecture organized across four maturity layers — governance, descriptive, diagnostic, prescriptive — so Sales, Marketing, and Finance can share a source of truth without sharing a surface.Healthcare GTM Analytics 09Platform Migration + ARR GrowthCross-functional migration and revenue expansion program spanning payments, systems, operations, and growth execution.Cross-Functional Delivery 10Revenue Recovery Audit WorkflowBilling validation workflow for mismatch detection, finance reconciliation, and recovered revenue.Enterprise Systems 11RMS Fare Validation SystemLogic-driven pricing validation model addressing complex fare rules and downstream revenue decisions.Decision Support 12Flight Disruption RecoveryDecision-support concept combining predictive signals, real-time data, and recovery visibility for irregular operations.Operational Decision Support

Hover a row.

§ 03Capabilities

Capabilities.

Where strategy, analytics, and AI meet — under operational pressure.

Product Strategy

Product direction when the problem isn't framed yet — finding the actual question, then writing the spec.

Business Analysis

A decade analyzing enterprise systems — actuarial pricing and risk modeling first, then billing, fare logic, and platform migrations. Always toward accuracy, auditability, and operational efficiency for operators.

Applied AI + Data

Forecasting layers, retrieval pipelines, and LLM-assisted decision tools. Independent builds where the AI sits inside the user's reasoning — designed for inputs you can audit and outputs you can defend.

§ 04About

I'm most at home in ambiguity, in the space before the structure exists.

A decade inside enterprise systems — three years actuarial analytics, then operator seats covering billing, pricing, fare logic, and platform migrations, where systems had to keep working under load. Recent independent work — Valuate, Tarazu, Sift, FocusForge, GridPulse — applies that approach to AI products.

Operating principle

I'll take the messy thing nobody wants to touch, break it apart, and make it work.

Product strategyBusiness analysisApplied AI

Diagnose first.

I reduce ambiguity early so teams can move with more confidence and less noise.

Design for trust.

If people don't actually use it, the work isn't done.

Ship the proof.

I care about what people actually use. The demo isn't the deliverable.

§ 05Contact

Open to roles, partnerships, and select consulting.

Currently exploring senior product roles and consulting work where systems thinking, analytics, and applied AI intersect — especially in operations-heavy contexts where decisions have to survive contact with the floor.

Email LinkedIn GitHub Resume