Vol. I · Decision Systems Lab

Building the Future of Human Decision Making.

I design intelligent systems that help people make better operational decisions. A working notebook at the intersection of AI, operations research, analytics, and human judgment.

01
Discipline
Operations Research
02
Method
Systems Engineering
03
Medium
AI + Analytics
04
Aim
Human Judgment
Mission

Most operational decisions are still made on intuition, scattered spreadsheets, and yesterday's data. I build the decision systems that change that: engines that fuse AI, optimization, and the texture of human expertise into one calm interface.

AI
Models that learn from operational reality, not just clean training data.
OPERATIONS RESEARCH
Optimization and simulation as the spine, not the marketing.
Judgment
Interfaces that respect the expert on the other side of the screen.
Featured System · 001

VELA.

Award Winning Research Prototype
First Place Winner, Do Track
NVIDIA Spark Hack Seattle
August 2026

A voice first healthcare navigation system that compares real prices, reasons across coverage options, asks rather than guesses, and keeps consequential action behind explicit consent.

Local AIHealthcare NavigationPrice TransparencyConsent Architecture
FIG 001 · VELA — System Flow
01Capture
Voice requestParakeet ASR
02Reasoning
NemotronProcedure matchingClarifying question
03Evidence
Hospital price transparency recordsInsurance context
04Decision
ComparisonPath selection
05Consent
Explicit consentScoped to the action
06Action
Booking or enrollment handoffTimestamped receipt
07Response
Magpie TTS
Portfolio

Systems I'm building.

Open Questions

Questions I'm exploring.

These are the prompts that organize the work. They aren't rhetorical, each one has a research thread.

  1. Q01

    How do we design AI that augments operational judgment rather than replacing it?

  2. Q02

    What does optimization look like when the objective function is contested by humans in the loop?

  3. Q03

    Can a decision log become the training set for the next generation of internal AI?

  4. Q04

    Where does deterministic logic end and probabilistic reasoning begin in supply chains?

  5. Q05

    How do we build systems that are calm by default and loud only when it matters?

Journal

Field notes.

JULY 28, 2026

Every variance deserves a theory.

A financial result is not yet an explanation. The Financial Variance Command Center asks six bounded roles to validate the record, decompose the bridge, connect operating context, challenge the forecast, inspect risk, and prepare the brief. The arithmetic stays deterministic. The final judgment stays human.

Read note →
JULY 26, 2026

The interface should look like the decision it supports.

A finance tool should feel precise before it feels futuristic. I replaced the generic AI-dashboard aesthetic with ledger structure, tabular numerals, institutional green, restrained risk color, and the visual hierarchy of an investment memorandum. Aesthetic is part of trust, but it cannot substitute for evidence.

Read note →
JULY 23, 2026

A good agent is a narrow promise.

The moment an agent can do everything, it becomes difficult to inspect anything. I started defining each agent by one bounded responsibility, one evidence surface, and one kind of claim. The goal is not simulated teamwork. It is attributable reasoning.

Read note →