Applied AI · Product · Enterprise Systems

Josh Lowry

I build production AI systems for regulated enterprises, align the people around them, and write the operating playbook when none exists.

10+years in regulated financial services
5+years shipping production AI
5K+documents processed annually
40%AI infrastructure cost reduction

Professional Overview

Applied AI and Enterprise Systems Builder with 10+ years of experience driving technical execution in regulated financial services, including 5+ years shipping production AI systems across a $46B+ asset manager, a Fortune 500 insurer, and startup environments. Deep expertise in production LLM systems, enterprise integrations, and cloud/data architecture, with a particular strength in aligning challenging stakeholders, modernizing legacy systems, and writing the rulebook where none exists. Built an AI document-processing platform handling 5,000+ documents annually at 95%+ accuracy, reducing turnaround from days to minutes and eliminating the need for a dedicated data-entry position. Also reduced AI infrastructure costs by 40% through intelligent model routing and previously scaled an underwriting platform to 5,000+ users while cutting quote turnaround by 60%.

Professional Experience

Clark Capital Management Group — Philadelphia, PA

$46B+ asset manager and Raymond James Investment Management boutique; acquired May 2026 while retaining its brand and leadership.

Director of Product, Technology

August 2026 — Present
  • Own AI product and technology strategy, translating executive priorities and complex advisor workflows into roadmaps, governance, and production solutions.
  • Lead AI provider evaluation and intelligent model routing, reducing infrastructure costs 40% while maintaining accuracy thresholds.
  • Drive productization of internal AI capabilities into broker-facing offerings with repeatable evaluation, accountability, and adoption practices.

Senior Systems Analyst

August 2023 — August 2026
  • Built an AI document-processing platform using Claude, Gemini, MLflow, and Databricks to handle 5,000+ documents annually at 95%+ accuracy, reducing turnaround from days to minutes and eliminating the need for a dedicated data-entry position.
  • Engineered Databricks, PySpark, and Snowflake pipelines operating at gigabyte-to-terabyte scale for 3,000+ clients.
  • Built an AWS serverless integration platform connecting Calendly, Microsoft Graph, and Salesforce across 60+ business units with 99.9% uptime and sub-second latency.

CORE STACK — Python · SQL · Databricks · PySpark · Snowflake · AWS · MLflow · Claude API · Gemini API · Salesforce API

Chubb Insurance — Philadelphia, PA

Fortune 500 global property and casualty insurer.

Associate Director

September 2021 — 2023
  • Led North America post-bind tools and actuarial models across four commercial lines; managed a 16-person team of six employees and ten contractors and achieved 85%+ adoption.
  • Aligned executive, underwriting, actuarial, and engineering stakeholders around strategy, delivery priorities, and measurable KPIs.

Senior Systems Analyst

December 2020 — August 2021
  • Built a custom automotive actuarial model for S&P 500 casualty clients using React, Python, R, Qlik, and Power BI.
  • Led analytics delivery and trained 20+ early-career associates.

Systems Analyst

January 2016 — November 2020
  • Scaled a company-wide underwriting platform to 5,000+ users, cutting quote turnaround 60%.
  • Automated audit reporting and data preparation, reducing manual workloads 80% across the review cycle.
  • Partnered with 250+ underwriters on requirements, UAT, training, and production support.

CORE STACK — Python · SQL · React · R · Oracle · Power BI · Qlik · VBA

Evergreen Eagle Capital

Research Analyst

January 2014 — December 2015
  • Produced public and private equity research for portfolio managers and built VBA/Excel tools for risk assessment and performance reporting.

Projects

COINAILYZER AI coin-grading app case study ↗

iOS and Android AI coin-grading application with guided image capture, on-device YOLOv8 Nano detection, TensorFlow Lite inference, explainability workflows, subscriptions, and receipt validation. Built against a 10,000+ image corpus with a domain partner.

MirrorMirror AI beauty platform case study ↗

Privacy-conscious AI beauty recommendation platform combining 270+ facial landmarks, computer vision, Llama, and retrieval-augmented generation to deliver personalized, explainable recommendations.