I work side by side with customer and engineering teams to turn difficult business problems into production-ready AI, data, and cloud software.
Across 15+ years in technology, I have moved from discovery and architecture into hands-on full-stack implementation, iterative validation, and production rollout in complex enterprise environments. I build the services, data paths, user workflows, and infrastructure around applied AI, then turn successful delivery patterns into reusable foundations for the next engagement.
I embed with customer, domain, and engineering teams to understand the problem in detail, build and test the right solution, and carry it through to production. I feed delivery lessons back into reusable architectures, tooling, and engineering practices.
Customer-facing, hands-on technical lead across discovery, solution design, full-stack implementation, evaluation, and production rollout. I translate ambiguous requirements into testable increments, code with customer teams in their environments, and align domain experts and engineers around delivery.
Worked from an ambiguous industrial operations problem to architect and build an agent-driven system that turns previously unseen machine logs into production-ready OpenTelemetry metrics and events, links them to digital-twin components, and routes approved signals into the operational data platform.
Designed the multi-agent workflow with custom MCP servers for access to tools and operational context, A2A communication for collaboration across specialized agents, multiple agent-memory technologies for durable and task-specific context, and reusable agent skills for repeatable domain operations.
The collaborating agents form an end-to-end conversational loop that decomposes raw log templates and extracts the operational information hidden inside them. They share context, ground interpretations in domain knowledge, validate facts, challenge intermediate results, and revise outputs until deterministic quality checks are met.
Built a fully automated knowledge-capture workflow in which an agent conducts structured interactions with domain experts, extracts and validates their operational knowledge, and persists it in a searchable knowledge base that grounds future agent decisions.
Iterated with domain and engineering teams while building the surrounding Python services, knowledge capture and retrieval, vector search, messaging workflows, relational persistence, deployment infrastructure, and review experience required to operate the system in production while retaining human review for ambiguous results.
Python · Microsoft Agent Framework · Custom MCP servers · A2A communication · Agent memory · Agent skills · Azure AI Foundry · Azure AI Search · Service Bus · Container Apps · SQL · Azure Data Explorer · OpenTelemetry
Owned the full data, model, and application path for an AI assistant that explains autonomous-driving behaviour from real vehicle context. Trained and fine-tuned multiple specialized models on vehicle-stack and visual data to produce targeted, context-aware responses, including an on-device intent-recognition model for constrained request routing.
Built the orchestration and backend services that combine those model outputs and route driver requests between grounded driving explanations, vehicle knowledge, and command APIs. Evaluated and iterated on model behaviour so responses matched the desired in-vehicle experience across representative driving situations.
Built a custom LLM evaluation framework using Microsoft AI Evaluation to measure model quality against representative driving scenarios and ground deployment decisions in observed performance rather than subjective assessment.
Developed in-car Python libraries that transform raw sensor, perception, and vehicle-state data into model-specific representations. Designed and implemented the end-to-end ingestion pipeline that captures event context, transfers it from the vehicle, stages it for low-latency retrieval, and supplies validated, model-ready inputs to the inference service.
Python · Multi-model training & fine-tuning · Microsoft AI Evaluation · Custom LLM evaluation · Vehicle-stack data · Visual data · Behaviour evaluation · Constrained intent routing · Sensor-data engineering · Protocol Buffers · Azure Cache for Redis · Container Apps · Cosmos DB
Built and iteratively refined a retrieval-augmented system for analyzing and comparing complex contracts. Designed hybrid-search indexes and ingestion pipelines, automated document extraction and metadata enrichment, and evolved prompt flows around real user question patterns.
Azure OpenAI · AI Search · Document Intelligence · Prompt Flow · Python
Helped establish a company-wide standard for core customer data by translating cross-team requirements into a reusable production foundation. Built APIs and serverless applications, test and performance tooling, and the landing-zone infrastructure for reliable multi-stage deployments without production disruption.
C# · ASP.NET · Azure Functions · AKS · SQL · Bicep · Azure Landing Zones
Led the development workstream for global rollout of an image-recognition system at self-checkout terminals. Productized the edge deployment, established CI/CD, and enabled independently versioned releases per store.
Azure edge devices · AKS · Helm · Container Registry · Private networking
Codified lessons from individual delivery teams into a reusable Azure blueprint for project-based data products. Extended the infrastructure codebase and built Python APIs that gave teams a consistent way to request and deploy services while improving delivery speed and alignment across the wider program.
Python · Event Hubs · Data Lake · Azure Functions · Pulumi · Bicep · Azure DevOps
Customer engineering lead for Azure adoption, modern device management, and long-term enterprise support.
Responsibilities
Impact: Improved operational maturity for managed workplace platforms and enabled stronger cross-customer knowledge sharing through a regional expert community.
Hands-on enterprise engineer for Configuration Manager, support delivery, and technical enablement.
Responsibilities
Impact: Increased customer platform stability and confidence through hands-on engineering support and structured consulting engagements.
Global workplace consultant spanning end-user service strategy, rollout automation, and client infrastructure operations.
Responsibilities
Impact: Helped scale and standardize workplace services globally, improving rollout consistency, operational control, and service quality.
