Tools I've built, programs I've led, apps I've shipped.

Legacy Data-Provider Transition

The office's investment models had run on a single market-data provider for more than 20 years. When something in the data looked off, I followed up and learned the provider was shutting down, with six weeks' notice. I negotiated an extension to several months and led the work.

• Ran demos with replacement providers and helped choose LSEG, then met with the outgoing provider's remaining staff to capture as much of their methodology as I could before they closed.
• Reverse-engineered the provider's discontinued proprietary calculations and rebuilt them on the new data.
• Converted the Excel model workbooks the investment process runs on, generating stock workbooks from a master template and batch-processing hundreds of files per run, with Power Query consolidation and automation that handles add-ins, link refreshes, calculation timing, and missing values.
• The models kept running through the cutover, and the automation cut prep time about 90% on the affected recurring reports.

Investment Research Database & Stock Dashboard

The investment team's stock research lives in ~3,900 Excel model workbooks, each built for depth on a single name. This added a cross-sectional layer on top: a queryable database and a browsable dashboard.

• A Python ingest reads every workbook into a SQLite database, with screening, cross-sectional, and ticker-history functions that surface names that may warrant a fresh look.
• A companion React dashboard adds search, company overviews, side-by-side comparison, sector filters, and relative-positioning charts.
• A companion stock-file database consolidates about 3,000 individual stock spreadsheets into one. Before cutover it was checked with a verification script and a parallel “shadow run” against the old files, which caught a bug that was silently dropping correction rows.

Daily Pipelines & Operations Automation

Scheduled tools that took recurring manual work off the investment team: about 40–50 hours a month across the cash report, trade checks, and earnings updates.

• Daily cash report: about an hour by hand each day, now about 30 seconds, with more detail, consistent formatting, and backfill for missed days.
• Trade-blotter cash checks: roughly two hours each by hand, several times a month; now about two minutes, and more thorough.
• Daily discount-rate model: recreates a rate the legacy provider used to supply, a core input to the office's models, from public market data. Runs every business day, logs history, backfills gaps, and emails the result to the team.
• Reconciliation tool: built for a trust team outside Investments; a quarterly reconciliation that took days by hand now runs in seconds.
• Shared plumbing: one email sender, a central path module that let every tool survive the move to SharePoint, and bulk-rename and account-matching utilities with preview and rollback.

Earnings Automation Suite

Three tools that replaced manual earnings data entry and a manual weekly scan.

• Earnings-trend updater: pulls earnings, dividend, and estimate data from LSEG into the stock-model workbooks through COM and recalculates valuation fields. It saves about 30 hours every quarter, and a one-time catch-up run brought long-stale names current.
• Weekly earnings alert: scans about 3,000 securities each week and emails beats, misses, revenue surprises, and outlook changes, with a built-in error alert.
• Stock-split alert: flags true splits and filters out depositary-receipt noise.

Portfolio Analytics: Concentration, Drift & Optimization

Working prototypes built on a reusable Python package I wrote for querying account data.

• Concentration monitor (prototype): flags single-stock concentration against target equity at the book, family, and account level, with dollar buy and sell labels, and compares sector exposure with the S&P 500. Test output never exposes client identifiers.
• Model-drift analysis (prototype): compares holdings with model positions to produce buy/sell drift signals, built to replace a manual monthly Excel workflow and matching its output format.
• Optimizer and risk summary (proofs of concept): mean-variance and Black–Litterman optimization across about 13 asset classes, and a one-page risk summary covering volatility, Sharpe and Sortino ratios, tracking error, capture ratios, and drawdown.

38-Year Asset Allocation Study

An allocation study over ~38 years of market data across U.S. equities, international equities, and intermediate fixed income, built as an interactive Plotly dashboard for use in client conversations.

• Efficient-frontier portfolios, rolling-period analysis, and how the optimal mix shifts over time.
• 60/40 and 70/30 versus all-equity, and growth of a hypothetical investment.

AI Wednesday

A weekly internal AI briefing written for finance and wealth-management professionals rather than technical specialists.

• Each edition covers industry developments, AI in finance and wealth management, new models and tools, risks, and a term of the week — with sources and plain-language explanations.
• Fixed editorial structure with source attribution and a Term of the Week, produced through an automated workflow; 22 editions as of October 2026.

AI Task Force & Enterprise AI Pilot

Member of Argent's AI Task Force since June 2026, and one of the most active participants in the firm's enterprise AI pilot.

• Recurring feedback to senior leadership on adoption, backed by a check-in workflow I built to track use cases and time saved.
• Documented and presented 18 automation and AI projects to senior leadership.
• Helped shape a documentation framework for internal tools.

AI Basics — Executives of Tulsa & Argent Lunch & Learn

• Executives of Tulsa: invited talk to about 25 retired executives on how large language models work, where they fail, and how non-technical users can start, with a one-page take-home guide. The Q&A ran 15 minutes over, and I was invited back.
• Argent Tulsa Lunch & Learn (May 2026): co-developed and co-presented an AI Basics session for ~20 employees, adapted for a trust-company audience with practical use cases and acceptable-use guidance.

Iron Ledger

Strength-training and nutrition tracker with a lift mode and an activity mode, used daily by me and a small group of friends.

• Custom, user-editable training programs stored in the database with friend sharing; friends graph with per-metric stat sharing and leaderboards.
• Recipes, meal slots, supplements, macro targets, weekly-average weigh-ins, Apple Watch workout import via HealthKit.
• Postgres with row-level security and server-side RPCs; native push; guided tour and What's New; 700+ automated tests. Native watchOS companion in progress.

Manifest

Offline-first planning app for weekly lists, packing lists, recipes, and grocery runs.

• IndexedDB mirror with an outbox for offline edits; email-OTP auth with row-level security; push reminders and scheduled notifications.
• Global search, locked/private lists, photo attachments, rich text, recipe import from URL, calendar view, onboarding tour and changelog.
• A Python archiver mirrors data nightly into Markdown notes with YAML frontmatter, with fixture-based tests. MVP plus ~50 post-MVP work orders shipped in roughly three weeks.

Quartermaster

Home inventory system for physical storage with printed QR labels that resolve straight to the item.

• Server-rendered Next.js with server actions and API routes; OTP auth and row-level security.
• AI-assisted itemization from photos, item types and links, consumables with low-stock tracking, maintenance reminders, multi-photo support.
• Crash observability, guided tour and help centre, ~20 verification scripts, nightly backup with restore drills.

Shell Oil Company / OU Corporate Valuation Challenge — Finalist

Produced an investment recommendation for a publicly traded corporation using comparable-company, precedent-transaction, and discounted-cash-flow models, presented to a panel of professional judges.