AI & machine learning, shipped — without the team overhead.
Leopard Data is a senior consultancy focused on AI, machine learning, and Kubernetes for enterprise clients — applied LLMs and AI agents, distributed ML, and production cloud — backed by 20+ years of shipping for Microsoft, IBM, Koch, Verizon, American Airlines, and the rest of the Fortune 500, all from a single architect with a sharp focus.
Why Leopard Data
Senior-architect focus. No hand-offs, no body shop, no slide-deck consultants.
AI & ML Engineering
Applied LLMs and AI agents on Claude, plus distributed machine learning (LSTM/DeepAR on Ray & KEDA) and ML.NET in production — and we know exactly when to override them.
One Senior Architect
Every line of code is reviewed by the architect who designed it. No knowledge gaps, no "let me check with the team," no junior developers learning on your dollar.
20+ Years of Production Delivery
From the Visual C++ compiler team at Microsoft in 1994 to AI-first SaaS in 2026. Reliable shipping for Fortune 500 clients across healthcare, finance, telecom, energy, and aviation.
Where We Go Deepest
Three disciplines we promote hardest — and have shipped to production for enterprise clients.
Clusters that run live systems
Production Kubernetes with Flux GitOps, Talos, EKS / AKS / OpenShift, and full observability — promoted across environments and engineered to stay up when it matters.
Explore Kubernetes AI & Machine LearningAI and ML that ship, not demo
LLM products and AI agents on Claude — RAG, MCP, agentic migration — plus distributed deep learning with LSTM / DeepAR, parallelized across Kubernetes with Ray and KEDA.
Explore AI & ML Cloud ArchitectureHyperscaler modernization
Azure, AWS, and Google Cloud — event-driven architecture, infrastructure-as-code with Pulumi and Terraform, and cost-disciplined, multi-region resilience.
Explore CloudWhat We Build
Six focus areas. All shipped to production, all backed by 20+ years of judgment.
AI-First Solutions Architecture
Claude integration, AI agents, RAG pipelines, MCP server design, structured tool use, token-aware cost control — including AI-driven code migration.
Machine Learning at Scale
PyTorch, ML.NET, TensorFlow, scikit-learn. LSTM/DeepAR forecasting and feature ranking, distributed across Kubernetes with Ray and KEDA.
Kubernetes & Microservices
AKS, EKS, GKE — Helm, ArgoCD, KEDA autoscaling, GitOps, zero-downtime rollouts. Multi-cluster when you need it.
Hyperscaler Cloud Modernization
Production deployments on Azure, AWS, and Google Cloud. Cost discipline, security architecture, multi-region resilience.
Enterprise .NET / Go / Python
Distributed systems, APIs, data pipelines, and durable functions. Battle-tested across financial services, healthcare, and telecom.
Infrastructure as Code
Pulumi and Terraform for reproducible Azure / AWS / GCP environments. Deploy two complete environments from one codebase.
The Stack We Ship With
Cut to what matters: every tool here appears in a case study or shipped project on this site.
AI & Machine Learning
AI in the Cloud — only the managed AI services used in production
Hyperscaler Cloud
Big Data & Data Engineering
Languages — 30+ across a 20+ year career; these are the daily drivers
App Frameworks
DevOps & CI/CD
Featured Project
One project we built in-house to prove the AI-first thesis.
Grade My Investments (GMI)
AI-Powered Stock Analysis & Rating Platform — full-stack SaaS on Azure, with Blazor web, .NET MAUI mobile, and a .NET 10 backend orchestrating three Claude models, ML.NET projections, and Stripe billing in production.
Client Case Studies
Delivery stories from the field.
AI Agents Migrating a Healthcare Platform from AWS to GCP
Built the AI agent pipeline — Claude on AWS Bedrock plus custom MCP servers — that analyzes and ports large FHIR/HL7 TypeScript codebases to Google Cloud, with the eventing, identity, and IaC to receive them.
Read the case studyDistributed Machine Learning at Scale
A feature-ranking engine running billions of calculations over 200K+ features, and an LSTM/DeepAR prediction engine — both parallelized across Kubernetes with KEDA and Ray.
Read the case studyKoch Industries — AI Production Forecasting
Distributed AI / ML platform with REST services and parallelized feature processing on AWS EKS, forecasting production capacity across multiple Koch business units.
Read the case studyTGL Golf — Game Operations Platform
CI/CD pipeline and Red Hat OpenShift Kubernetes platform that deploys all 16 game components onto the on-prem cluster running the indoor golf league on game day at the Florida arena.
Read the case studySpectocor — Healthcare Analytics Recovery
Diagnosed and rescued a stalled $15M heart-analytics platform investment. Replaced a 160-developer offshore program with five senior onshore engineers, personally interviewed and hired before handoff.
Read the case studyFiserv — Always-On Digital Banking at 10M-User Scale
Top technologist on a 60+ person program building a cloud stand-in digital banking system on Azure — Cosmos DB, KEDA, and regulator-grade PII security for 2,000+ banks and 10 million banking customers.
Read the case studyFM Global — Property Risk Analytics Re-Architecture
A year as Solutions Architect at one of the world's largest commercial property insurers — geospatial data architecture, a KEDA/Kafka autoscaling PoC, Azure Data Factory adoption, and ownership of application security.
Read the case studyCentene — .NET Modernization & DevSecOps at Scale
Authored the .NET 8 upgrade playbook multiple teams followed at a Fortune-25 managed-care enterprise, hardened GitLab CI/CD and container security, and was named lead for the platform's AWS migration.
Read the case studyNational Oilwell Varco — Go Microservices at the Drilling Edge
Go microservices with gRPC streaming running on edge devices and in the cloud — NATS/MQTT/Kafka data sync, SAML/Okta security architecture, and scrum-master leadership of the product team.
Read the case studyGreenfield Multi-Tenant Geospatial SaaS
Solutions Architect from RFP to running platform for a national subsurface-mapping company — multi-tenant schema, KML/KMZ/GeoJSON ingestion, ~30-endpoint map services on AKS, and ERP/billing integrations.
Read the case studyWindows-to-Linux Containerization for a National Telecom
Took Windows-only .NET OSS services to Linux containers — isolating Windows dependencies, solving DB2/Informix native clients on Linux with IBM, and publishing the playbook other teams followed.
Read the case studySoci — B2B Price Catalog Distribution
EDI 832 price catalogs published from NetSuite straight into dealers' purchasing systems — designed, built, hosted on Google Cloud, and operated for eight years.
Read the case study
W. Allan Edwards
Allan started writing software professionally at 19, recruited to Microsoft to work on the Visual C++ compiler team for the PowerPC architecture. He was hand-picked for one of Bill Gates' pet projects — Visual Basic for Windows CE — where he wrote portions of the runtime in C.
That was 1994. Decades later he's still shipping production code, now AI-first, for clients like IBM, Koch Industries, Fiserv, FM Global, Bell Helicopter, Thomson Reuters, Pioneer Natural Resources, National Oilwell Varco, SiriusXM, Verizon, and American Airlines.
Allan is the architect, lead engineer, and proprietor of Leopard Data. He works exclusively corp-to-corp from Plano, TX.
Trusted By
A short list of teams Leopard Data has shipped for over two decades.
Allan is an exceptional software engineer that can pick up the barest of requirements and knock it out of the park. He was able to navigate through several scrum teams and Dev environments to deploy enterprise analytics solutions across multiple sites and native apps that our in-house team struggled with. If you’re looking for someone to get things done right and fast, Allan is your man.
Have a hard problem? Let’s build it.
Senior architecture, hands-on engineering, AI-first delivery. Corp-to-Corp engagements out of Plano, TX — available now.