krishna dutt

• Streamlined Fin-ops reporting for 8k+ pods clusters by leveraging Kubecost, resulting in 25% cost reduction. • Enhanced observability platform by metric Governance by coordinating with teams four time zones and improved alerting thereby reducing MTTR by 40%. • Migrated in-house observability platform to New-Relic, thereby reducing mean time to detect (MTTD) by 30%. • Developed multi-media batch processing in Ruby and Clojure reducing processing time by 10%.

• Administered SAAS platform on 5 On-prem Kubernetes clusters and deployed MLOps tools such as MLFLow, Airflow, Deltalake etc. • Implemented Proof of Authority based Private Ethereum blockchain for creation of NFT with for 4 million athletes. • Deployed a self-Managing Kubernetes cluster in On-Prem and administered around 90 pods with EFK stack. • Architected a novel Event-driven Server-less Monitoring system, using Terraform and AWS, which decreased reaction time by 90 % and cost by 70 %.

• Lead and delivered a methodology to predict above ground biomass (AGB) using Deep Learning CNN framework in AWS (S3 buckets, Lambda triggers and sage maker), which performed 15% better than existing state- of- the - art. • Worked in conjunction with European Space Station (ESA) phi labs, using sentinel 2 satellite data, over the southern Italian region, comprising of 30 million data points streamed in every 5-day interval.

• Authored a Climate Prediction model for Milan, based on time-series Regional Climatic Model (RCM), WUDAPT model, ML models, with 95.5% accuracy, which is used to choose economically viable solution to improve climate resilience of the local area

• Automated data collection process, which saved 90% time and hosted about 20M PostgreSQL via Geo-server. • Computed Multi variable Regression models for future energy demand, Environmental risk Assessments for a population of 1,00,000 based on the development trajectory using geo-spatial satellite data of Landsat-8, Aster etc.