Systems · Data · Responsible AI
Research & Innovation
From reliable distributed platforms to data-centric intelligence—exploring ideas that scale in production and stay understandable for the teams who operate them.
Research areas
AI & machine learning
Applied ML for structured decision support, evaluation under drift, and patterns that keep models observable and maintainable in real pipelines.
Data science & analytics
End-to-end analytics workflows, feature quality, and communication of uncertainty from raw signals to stakeholder-ready insight.
Distributed systems
Resilience, scalability, and operability—designing services and data paths that degrade gracefully under load and change.
Software & platforms
Developer experience, release safety, and platform ergonomics so research prototypes can mature without losing clarity.
Current projects
Focus
Ongoing
Funded
Reliable pipelines for streaming features
Contracts, backfill strategy, and monitoring so feature stores stay trustworthy as schemas and sources evolve.
Ongoing
Evaluation loops for production ML
Lightweight harnesses for regression detection, shadow traffic, and human-in-the-loop review hooks.
Ongoing
Collaboration
Human-readable analytics for operators
Dashboards and narratives that connect anomalies to likely causes—designed with on-call engineers in mind.
Past projects
Completed
Edge-aware sync for field devices
Conflict resolution and bandwidth-aware replication for intermittently connected clients; outcomes fed a product playbook.
Completed
Curriculum lab for data engineering
Open modules for ELT patterns and testing data jobs—adopted in multiple cohorts and workshop series.
Completed
Benchmark suite for retrieval quality
Reproducible tasks and metrics for semantic search over internal knowledge bases; released as an internal standard.
Research timeline
-
2025
Production ML observability initiative
Launched cross-team standards for tracing model inputs, outputs, and drift alerts in shared services.
-
2024
Data pipeline reliability program
Defined SLOs for batch and streaming jobs; rolled out playbooks adopted by platform and product squads.
-
2023
Distributed systems reading group → lab practices
Translated classic papers into internal tech talks and mentoring tracks for junior researchers.
-
2022
First cross-venue publication push
Aligned empirical systems work with archival venues and open artifacts for reproducibility.
Research interests
Machine learning
Distributed systems
Data engineering
MLOps
Reliability
Stream processing
Software architecture
Human–computer interaction
Knowledge graphs
Information retrieval
Cloud security
Observability
Open source
Reproducibility