Farrukh Nauman

◦ Lead a 4–7 person DS/analytics team; own roadmap, stakeholder steering, and technical direction across telemetry analytics, ML, and GenAI automation.

◦ Coordinate transition planning/execution for Databricks → Snowflake while maintaining daily Spark/BI workloads with several terabytes of sensor data.

◦ Built a cross-platform migration validation system using AI coding agents: automated comparison of schemas, row counts, key distributions, and date ranges across dozens of tables—from zero to live dashboard in ∼1 week.

◦ Connected Databricks and Snowflake CLIs through an autonomous coding agent (Codex CLI / Cortex Code CLI) to orchestrate migration, validation, and reconciliation pipelines end-to-end—shrinking multi-day manual runbooks

into hands-off workflows.

◦ Redesigned a critical daily pipeline from full recompute to validated incremental processing: 5–10× speedup with row-level correctness proof, projected to cut compute costs by 50%+. Killed a faster but fundamentally broken optimization using multi-scale benchmarks and EXCEPT-based validation.

◦ Open-sourced ts-agents (PoC): a framework for autonomous time series agents; applied it to machine activity recognition from sensor data, demonstrating that classifier-heavy activity recognition workflows can be replaced by agent-driven analysis.

◦ Prototyped Text-to-SQL workflow automation (PoC): synthetic evaluation dataset generation + instrumentation/logging to analyze failure modes.

◦ Developed a practitioner framework for large code migrations with coding agents: evidence-preservation methodology, behavioral oracles, and AGENTS.md-based quality contracts.

Project Lead: Sustainable Fashion AI Automation (2022-2025: 24 months): Leading two major initiatives in sustainable fashion

• Challenge: Helsingborg school district lacked ability to identify at-risk students early, resulting in up to 40% failure rate in some schools in 9th grade.

• Solution: Developed predictive analytics system identifying absence, poor grades in English and Math as the key indicators in 6th grade that predict 9th grade performance, with school-level feature analysis for targeted funding.

• Impact: Enabled early intervention for 10% of the student population, and provided data-driven policy recommendations impacting 3,000+ students.

• Technologies: Azure DevOps, Azure Functions