Design and own high-throughput, low-latency ML systems (2000+ RPS) for TravelAds, including multi-service training and serving architectures, auction and ranking models, and real-time inference services that meet strict sub-100ms SLAs.

Build and evolve ML infrastructure and data foundations – feature stores, online/offline feature pipelines, embedding and vector services, and data lineage and versioning – that power ad relevance, bidding optimization, experimentation, and model evaluation at scale.

Accelerate the end-to-end ML lifecycle by automating training, validation, deployment, shadow testing, A/B testing, and retraining using orchestrated workflows (e.g., Flyte, Airflow) and robust quality gates.

Develop agentic AI and LLM/RAG-powered workflows that automate ML operations (training, deployment, validation, monitoring, calibration) and enable AI-assisted dataset creation, operational analysis, and decision support.

Define and implement ML observability, reliability, and cost guardrails through drift and feature-freshness monitoring, health dashboards, SLO/SLI definitions, incident response, and resilience-focused improvements.

Safely integrates and operates AI/ML-enabled solutions that improve outcomes, while setting technical direction, mentoring MLEs to operate independently, and leading cross-team initiatives that elevate ML engineering practices and business impact.