Skills Summary¶
Condensed from the CV
Drafted with AI assistance, checked line by line against the full record. Last verified 2026-09-09.
Seventeen years in enterprise data platform delivery. The full record is on the CV; this is the scannable version.
Technical¶
| Area | Detail |
|---|---|
| Cloud | AWS, GCP, Azure — multi-cloud architecture and migration |
| Data platforms | Snowflake, Databricks, Airflow, Spark, Kafka, Delta Lake, Unity Catalog |
| AI and ML | Agentic RAG, retrieval architecture, LLM-as-judge evaluation, MLflow, DSPy, BERT/CoreNLP on SageMaker |
| Infrastructure | Terraform, AWS CDK, Kubernetes, service mesh (HAProxy, Linkerd), GitLab CI/CD |
| Languages | Python, SQL, TypeScript; full-stack .NET background |
| Analytics | Power BI / DAX, ThoughtSpot, dimensional modelling, semantic layers |
Delivery¶
- Distributed teams from 3 to 70 people, in Agile and SAFe environments.
- Multi-stream programme coordination — a 10-squad editorial migration, a 25-person retail DWH modernisation.
- Stakeholder management across matrixed and regulated organisations.
- Staffing, performance management, conflict resolution, budget forecasting.
- Measured outcomes: team velocity doubled on one engagement; delivery speed up more than 30% on another; a client's roadmap revised after delivery ran "twice as fast as originally expected".
Engineering excellence¶
- SDLC, CI/CD and docs-as-code established from scratch on multiple greenfield engagements.
- SAFe adoption, SIPOC process mapping, ceremony redesign.
- Pipeline performance improvements up to 90%+ — one orchestrated pipeline taken from 4 hours to 15 minutes.
- Manual effort reduced 50–60% through automation and reconciliation tooling.
Generative AI¶
- Agentic Code Converter for SQL migration (Impala → Athena/Snowflake, Snowflake → dbt Core) reaching 60% automation coverage, built on EPAM's migVisor suite with custom prompt engineering.
- Governed knowledge platform for a global CPG manufacturer — ingestion through evaluation, described in Work.
- NLP risk extraction for financial crime signals — BERT and CoreNLP on AWS SageMaker, delivered from scratch.
- Second place, EPAM Data Analytics Hackathon, November 2025 — MCP and A2A integration for migVisor.
Most recent large programme¶
Data platform migration tooling, March – September 2025. Led an 8-person team enabling 70+ Data Product Engineering teams to move 500TB+ across 200+ applications from Cloudera (Impala/Kudu) to an AWS lakehouse. Three tools delivered in parallel: historical data migration (Oozie/Spark/AWS DataSync), the AI-powered code converter above, and reconciliation tooling that halved manual verification — while the programme's scope evolved from lift-and-shift to full replatforming mid-flight.