Rajesh Dhanda

Led MLOps & Agentic-RAG foundations for DS and GenAI services on Azure & Databricks across multiple product teams.

• Global Consumer Help Center: AI search across 15+ brand help centers & 400+ brand websites spanning multiple countries, markets & languages — reducing cost and reliance on human agents
• Procurement Contract Intelligence: AI-powered platform analyzing 11,000+ Contract IDs & 30,000+ contracts covering $7B third-party spend — enabling faster review, compliance visibility & time-cost savings
• Commercial Tech: MLOps framework for Marketing Mix Models, scalable, repeatable modeling of brand, market–level sales drivers - used to optimize global marketing spend
• MLOps and Agentic-RAG Foundation: Framework with code & detailed docs, used across Help Center, Contract AI, NRM, NBA, QSC and Commercial Tech projects

Productionized DS, Data Engg, and BI services/pipelines on GCP (GKE, KServe, Triton, Airflow), implementing best practices, monitoring and cost optimization.

• ML Platform — Serving: Migrated 30+ DS workloads to GKE; Inference Store: EdgeDB with 5ms latency across 10+ services; Feature Store: FastAPI + BigQuery + Redis with 10ms latency
• Orchestration Platform: Airflow on GKE — 70+ git-synced DAGs, 1000+ daily runs, 3800+ task executions supporting DS-BI teams
• Data Flow Pipeline: Syncing Data 2500 Sheets with 15000 Jobs/day. (Refer to Detailed Blog got cited in Modern Data Stack and Data Engineering Weekly)
• NBFC DS Migration: AWS → GCP, monolith → GKE microservices, latency from 3s to 503ms (99.99%), eliminating 100% timeouts previously at 8–9%

Object detection & tracking (DeepSort), healthcare symptom mapping algorithms, TFLite on Raspberry Pi.

• Deep Sight AI Labs: Vehicle detection, tracking & counting using YOLOv4 + DeepSort on traffic surveillance footage
• DocBot+: Healthcare symptom-to-disease mapping algorithms, TFLite model deployment on Raspberry Pi for edge inference