Enterprise Data & AI Solution Lead at IBM's One Microsoft Practice, focused on Financial Services. I design and deliver applied AI, agentic systems, and real-time data architectures on Databricks & Microsoft Fabric. I turn complex requirements into production-grade solutions that win deals and drive measurable outcomes for banking, capital markets, and insurance.
Where I focus, and the point of view behind each area. Hover or tap a card to read the full take.
The model was never the hard part
Frontier models change every few weeks; the architecture should not. I design agentic systems where the model is a swappable component and the schemas, gates, and audit trail are the fixed part. The systems that reach production in a bank are not the most autonomous ones. They are the ones an examiner can inspect.
See use casesThe micro-batch era is ending
Extended trading hours and T+1 settlement turned a minutes-behind operational data store into a liability. With Databricks Real-Time Mode and Lakebase generally available, the real-time ODS is now an architecture decision, not a research project. I design the streaming path so intraday risk, position, and P&L are current the moment someone asks.
View architecturesThe "which one" war is over
At FabCon 2026, Microsoft and Databricks started dissolving the boundary between their platforms. The question stopped being which one to pick and became how to govern both. I design hybrid lakehouse architectures with a single governance layer across Delta, Unity Catalog, and OneLake, so engineers and analysts work from the same data without a copy war.
Platform deep diveTechnical clarity wins the deal
Most enterprise AI deals are not lost on price. They are lost on a buyer who cannot picture the system running in production. I lead technical discovery, reference architecture, and live demos that turn a vague ambition into something a risk committee and a CFO can both approve.
Discuss an opportunityThe unglamorous layer that decides everything
Agents and dashboards get the attention; the pipeline underneath decides whether either can be trusted. I build production data engineering for financial workloads: trade capture, market data ingestion, change data capture from core banking, and semantic models an examiner can trace back to source.
View pipelinesTrust is an architecture, not a policy
In a regulated institution, governance is not paperwork applied after the build. It is access, lineage, model risk, and now cost, designed in as first-class dimensions. As spend controls join access and audit inside the platform itself, the governed surface only grows. I design for the examiner from the first sprint.
See governance modelTwo platforms, one architecture
IBM's One Microsoft Practice and the Databricks partnership are most useful to a client when they arrive as one coherent architecture, not two vendor pitches. I drive joint solution design and co-build motions that keep the client's outcome ahead of either logo.
Explore partnershipAt the intersection of applied AI, data platforms, and financial services delivery.
I lead solution architecture, presales, and delivery for Data & AI in financial services inside IBM's One Microsoft Practice. The brief: turn complex regulatory and capital-markets requirements into production systems on Databricks and Microsoft Fabric, and win the work that gets there.
Validated expertise across data engineering, applied AI, and cloud platforms.
Advanced lakehouse design, optimization, and governance skills for production workloads on Delta Lake and Unity Catalog.
Hands-on proficiency with retrieval-augmented generation patterns, agentic architectures, and responsible LLM delivery.
Expertise in building resilient Spark applications tuned for scalability and performance at financial-grade SLAs.
Designing and orchestrating modern data platforms across Azure Synapse, Fabric, and related services for enterprise workloads.
Grounded understanding of Azure AI services, ethical considerations, and deployment best practices for regulated industries.
Recent initiatives delivering applied AI and real-time data solutions for financial services enterprises.
Multi-agent system that automates know-your-customer and anti-money-laundering document review. It extracts entities, cross-references watchlists, and generates risk assessments with full audit trails. Reduced manual review time by 60%.
Discuss this projectStreaming operational data store on Databricks + Fabric for real-time position aggregation, intraday VaR calculation, and P&L attribution. Delivers sub-second analytics to risk desks and front-office traders.
Discuss this projectAI-powered claims triage and adjuster copilot built on Azure OpenAI and Databricks Mosaic AI. Automates document intake, damage assessment, and settlement recommendations with compliance guardrails.
Discuss this projectInsights on applied AI, financial services technology, and the data platforms shaping 2025–2026.
An architect's point of view: walking a KYC agent from a notebook to production on the Databricks agentic stack, built with Agent Bricks and governed with Unity Catalog.
Read ArticleDatabricks put AI spend controls inside Unity AI Gateway. Why agentic AI cost is non-linear, why an unbounded cloud bill is a governance gap for banks, and how to make AI spend something a risk committee can approve.
Read ArticleAgentic AI is in production at major banks, and seven frontier models shipped in one quarter. Why the model is the least interesting decision, and compliance engineering is what actually decides who ships.
Read ArticleExtended 23-hour trading and T+1 settlement turned a minutes-behind operational data store into a liability, just as Databricks Real-Time Mode and Lakebase reached GA. Why 2026 forces the architecture question, and how to answer it.
Read ArticleAt FabCon 2026, Microsoft and Databricks began dissolving the boundary between their platforms. Why the Databricks-versus-Fabric question is settled, and why governance discipline is the real problem now.
Read ArticleA personal take on how Databricks RTM, IBM's $11B Confluent deal, and the next wave of ODS architecture fit together for capital markets. What's strategic, what's hype, and what I'm changing in my reference designs.
Read ArticleChoose the conversation that fits best. Each session includes prep materials and follow-up insights.