Patrick McDowell

AI Product Engineer | Applied AI | Decision Systems

London, United Kingdom · +44 7442 106062 pmcdowellthe3@gmail.com · patrickmcdowell.dev LinkedIn · GitHub

AI product engineer who builds AI and data products from ambiguous problem through prototype, architecture, implementation, and adoption. At Wellington Management, I build internal AI platforms and decision systems at a $1.4T asset manager, including a product used daily by 100+ investment professionals. I work directly with expert users to replace fragmented workflows with tools that become part of how they work.

01 Experience

Wellington Management

Investment Science & Risk Analyst

London · 2022-Present

Risk MCP / AI Product Platform

  • Designed and built an internal platform on the Model Context Protocol for AI-powered knowledge retrieval, dashboard generation, and reusable AI development workflows.
  • Enabled non-developers to create dashboards and analytical tools with AI, turning a one-person bottleneck into a self-service capability.
  • Rapidly prototype, test, and iterate new capabilities with investment professionals around real workflows and business problems.
  • Help shape the team's AI adoption strategy by prioritizing useful workflows over isolated demonstrations.

Equity Risk Explorer

  • Conceived, designed, and built the firm's flagship equity risk product, now used daily by 100+ portfolio managers, risk managers, client strategists, and senior leaders.
  • Unified holdings, factor exposures, attribution, historical analytics, scenarios, returns, and pre-trade analysis from 6+ internal systems into one workflow.
  • Built and continue to maintain thousands of lines of Python powering seven analytical modules designed around how investors actually analyze portfolios.
  • Expanded adoption beyond the original audience through direct user feedback and fast, iterative product development.

Decision Systems & Optimization

  • Built portfolio optimization and scenario-analysis tools supporting portfolio construction and investment decisions.
  • Designed and backtested a lower-risk optimized strategy that helped win an institutional client pitch and was adopted into client materials.

02 Selected Product Work

Full Court Press

fcp.patrickmcdowell.dev

  • Built a fantasy basketball decision platform end to end: a Python/FastAPI/Postgres backend that ingests ESPN league history, reconstructs daily rosters, matchups, box scores, and transactions, and powers draft, waiver, lineup, and season-planning tools.
  • Built and backtested a pickup recommender across 616 historical decisions from a 14-team league, worth +0.17 categories per matchup for streaming moves and +0.98 to +1.06 over 30 days for season moves, then deployed it on a VPS with scheduled ingestion, player-status monitoring, alerts, and a daily digest.

Sons & Daughters of the U.S. Middle Passage

sdusmp.patrickmcdowell.dev

  • Consulted with a national lineage society to rethink how AI could improve archive search, registrar review, and the workflows around them.
  • Designed a working product mockup with the organization that is now being implemented.

03 Earlier Experience

Senior Financial Analyst · Wellington Management

London · 2020-2022

  • Built self-service Tableau reporting used by senior leaders across EMEA and APAC, replacing fragmented Excel workflows and reducing a recurring monthly reporting cycle by roughly 40 analyst-hours.
  • Automated reporting pipelines with SQL and Python so executives could explore assets, revenue, net client growth, and regional performance on demand.

Fund Controller · Wellington Management

Boston · 2017-2020

  • Built Python tools that generated financial statement disclosures and automated workflows supporting 200+ investment funds.
  • Reduced financial statement review from roughly three hours to 30 minutes through automated validation; the tools were standardized across the team.

Financial Analyst · Arrowstreet Capital

Boston · 2015-2017

  • Wrote complex SQL for operational and business analytics and automated recurring reporting with SQL, Excel, and Python.

Audit Associate · KPMG

New Jersey · 2012-2015

  • Taught myself VBA to automate repetitive reconciliations and reduce manual copy-and-paste, establishing the habit of rebuilding a process instead of repeating it.

04 Skills

AILLMs, Model Context Protocol (MCP), prompt engineering, internal AI products, AI-assisted development

ProductProblem discovery, rapid prototyping, user research, workflow design, iterative development, product adoption

EngineeringPython, SQL, Pandas, NumPy, SciPy, Plotly, Git

Data & AnalyticsTableau, data visualization, optimization, scenario analysis, decision systems

05 Education

Rider University · B.S. in Accounting