Josh Lowry — Forward Deployed Engineer

5+Years Production
AI / LLM

95%Doc Processing
Accuracy

5K+Enterprise Users
Served

40%Infra Cost
Reduction

Experience

Senior Software Engineer

2023 — Present

Clark Capital Management Group · Philadelphia, PA

Embedded as the sole AI engineer inside a $47B AUM asset manager — scoped the platform, selected the models, designed the evaluation framework, and shipped end-to-end without a playbook or team. That's the 0→1 FDE motion.
Owned technical delivery from prototype to production for core AI infrastructure, balancing inference cost, latency, and evaluation quality in regulated financial-services workflows.
Managed technical discovery and scoping for ambiguous AI deployments, translating business constraints into production-grade systems, eval criteria, and rollout plans.
Built multi-model orchestration pipeline (Claude + Gemini) processing 5,000+ brokerage documents annually across 15+ custodians at 95%+ accuracy — turnaround from days to minutes, eliminating a manual FTE bottleneck.
Engineered serverless integration stack (Calendly · Microsoft Graph · Salesforce via AWS Lambda) automating advisor scheduling across 60+ business units — 99.9% uptime, sub-second latency.
Led AI provider evaluation and intelligent model routing; reduced infrastructure costs 40% while maintaining accuracy thresholds. Drove productization of internal AI tooling into external broker-facing offering.

Associate Director

2021 — 2023

Chubb Insurance · Philadelphia, PA

Led North America post-bind tooling and actuarial models across middle market ($1M+ revenue), major accounts ($1B+ revenue), healthcare, and international commercial lines; managed 16-person team.
Achieved 85%+ enterprise adoption across thousands of users — the gap between a working demo and an organization that depends on it is almost entirely a human problem. I learned to close that gap.
Built strategic initiatives with C-suite; drove alignment across underwriting, actuarial, and engineering in HIPAA-regulated and SOC 2-compliant environments.

Systems Analyst → Senior Systems Analyst

2016 — 2021

Chubb Insurance · Philadelphia, PA

Architected company-wide underwriting platform for 5,000+ employees — clearance, rating, pricing, and proposals. Cut quote turnaround 60%.
Reduced manual workloads 80% across the audit review cycle through automated data cleansing and compliance reporting. Built actuarial automotive model for S&P 500 casualty businesses.

Research Analyst

2014 — 2016

Evergreen Eagle Capital · Equity & Private Equity Research

Built financial automation tools for portfolio risk assessment and daily performance reporting, giving me early domain depth in investment workflows, data quality, and decision-support systems.

Also Building

MirrorMirror — AI skin analysis & makeup recommendations, published to App Store & Google Play.

COINAILYZER — React Native AI coin-grading app (iOS & Android). Proof of full-stack product ownership: real-time on-device YOLOv8 Nano TFLite detection, automated obverse/reverse capture, backend grading with polling and stale-response protection, subscriptions, and receipt validation. Built and shipped end-to-end. Deep dive →

The best way to understand a tool is to build something with it.

Selected Deployment

Brokerage Document Intelligence

Embedded with operations, compliance, and engineering stakeholders to turn a manual brokerage-document workflow into a production LLM system. Owned discovery, technical scoping, model evaluation, deployment, and rollout. Built a queue-based AWS Lambda pipeline with encrypted document storage, Snowflake metadata/result tables, model routing by document class, source-grounded extraction, exception queues, reviewer feedback loops, and field-level evals. Result: 5,000+ annual documents processed at 95%+ accuracy, turnaround reduced from days to minutes, and 40% lower inference cost through model routing.

The FDE motion — embedded with a customer, no playbook, delivery is urgent — is not something I'd be learning. It's what I've been doing inside financial services, where the data is sensitive and stakeholders know immediately when something fails.

I'm drawn to FDE because I want to stay close to the workflow: where the model has to meet data boundaries, user trust, review loops, production constraints, and measurable adoption.

I've shipped AI systems across the full deployment loop: discovery, scoping, model selection, evals, production rollout, stakeholder adoption, and post-launch iteration. My strongest pattern is turning messy regulated workflows into durable systems that balance cost, latency, quality, auditability, and user trust.

Clark Capital Management Group · 2023–Present

A decade of customer-embedded AI deployment in financial services. Asset management at CCMG ($47B AUM). Middle market through $1B+ major accounts, healthcare, and international lines at Chubb. Environments where the cost of a failed deployment isn't theoretical — it's a phone call from a portfolio manager or an underwriter.

The 0→1 enterprise AI deployment experience
OpenAI's FDE org is built around.

Stack

AI / LLM

Multi-Model Orch. Prompt Engineering Claude API Gemini API RAG Agent Dev MLflow

Integrations

Salesforce API Microsoft Graph Calendly Webhooks Box API

Cloud & Infra

AWS Lambda S3 · EventBridge SQS / SNS CloudWatch Playwright

Data

Databricks PySpark Snowflake Python SQL

Compliance & Regulatory

HIPAA SOC 2 Audit Trails Regulated Environments

Education

B.S. Accounting & Finance

Missouri Southern State University · 2014

Built For

End-to-end production AI deployment embedded with enterprise customers — no playbook, delivery is urgent, failure has real consequences.

Regulated industries where trust, reliability, and stakeholder scrutiny are table stakes, not edge cases.

The 0→1 motion: scoping, model selection, evaluation framework, shipping, and then making the case for what comes next.