Apr 2020 – Mar 2024 · 3 yrs 11 mos

• Built and scaled centralized cloud infrastructure and internal platforms supporting Application workloads
• and Analytics and MLOps workloads, with a focus on reliability, observability, and cost-efficient
• infrastructure automation for global engineering teams.
• Architected and deployed an Airflow Platform-as-a-Service (Terraform, ArgoCD, Helm) with
• custom RBAC, secrets backend, centralized logging, and observability; led migration from Google
• Cloud Composer, reducing platform costs by 70% while improving operational control and
• reliability
• Designed and implemented a highly available Prometheus observability stack integrated with
• PagerDuty and DeadMansSnitch, achieving 99.99% platform uptime and saving $500K annually
• by retiring Azure Container Insights
• Introduced spot-instance orchestration across Kubernetes and Databricks workloads, optimized
• GPU and compute utilization, and reduced annual infrastructure spend by $100K+
• Implemented governance and cost observability frameworks (Databricks Overwatch) to provide
• automated insights into platform inefficiencies and resource usage patterns

Apr 2020 – Apr 2022 · 2 yrs

• Working on developing Data Platform for Data and MLOps

Jul 2017 – Apr 2020 · 2 yrs 9 mos · Gurugram

• Part of the Financial Modeling Group’s Data Infrastructure initiative, focused on modernizing and automating big data systems for quantitative equity research and analytics.
• Architected a modular Data Fabric for Equity Research to automate ingestion and storage of multi-
• source structured and semi-structured datasets, enabling standardized signal generation workflows
• and reusable compute/analytics layers across research teams (GCP, Python, Flask, MongoDB,
• Ansible)
• Led migration of on-prem mortgage asset modeling infrastructure to GCP and Airflow (Composer),
• modernizing legacy batch pipelines, reducing runtime from 48 hours to 10 hours, and significantly
• improving research iteration cycles
• Designed and implemented a scalable Data Lake platform for low-latency interactive analytics,
• establishing governance and data organization patterns (Medallion-style layering) to prevent data
• swamp and support large-scale analytical workloads (Hadoop, Spark, Presto)
• Engineered performance-critical internal tooling (FTPSync) for distributed file system
• synchronization across HDFS, NFS, and object storage, reducing algorithmic complexity from
• O(n²) to O(n) and lowering memory footprint by 20%
• Impact: Accelerated research workflows and enabled scalable data processing for global investment teams.

Oct 2016 – Jul 2017 · 9 mos

• Worked on large-scale Big Data and Advanced Analytics systems for strategic enterprise and public-sector
• clients, focusing on distributed data pipelines, performance optimization, and scalable data infrastructure
• for analytics-driven decision making.
• Developed distributed financial fraud detection pipelines using Spark, Neo4j, and Python to identify
• fraud rings and shell entities; optimized pipeline architecture to reduce runtime from 12+ hours to
• under 2 hours for large graph based datasets
• Designed and implemented Hadoop-based Data Lake and ETL frameworks (Hive, Spark, Ranger,
• Avro) integrating structured and semi-structured data sources, enabling scalable analytics and
• warehousing on self-hosted Hortonworks clusters
• Engineered production-grade data processing workflows for high-volume analytical workloads,
• improving reliability, data consistency, and execution efficiency across client environments

Jan 2016 – Oct 2016 · 9 mos

Jan 2015 – Jan 2016 · 1 yr · New Delhi Area, India

• Worked on Machine Learning based Business solutions to provide actionable business insights. I worked on various problems such as user behaviour modelling, customer segmentation, sentiment analysis etc.
• Technology Stack: python, flask, pandas, scikits, numpy, MongoDB, elastic-search

Aug 2014 – Jan 2015 · 5 mos

• Worked on structured data mining, API development, implemented large scale distributed crawlers to collect information from various sources.
• Responsibilities also included direct client interaction, cost and business
• analysis, and end to end project delivery.
• Technology Stack: python, Google Compute Engine

May 2014 – Aug 2014 · 3 mos

• I worked on problems which included information retrieval, ranking and social network visualisation.
• Technology Stack- python, networkx

Aug 2012 – Apr 2014 · 1 yr 8 mos

• Responsibilities included substitute lectures, organising tutorials and practical lab sessions, lab viva, evaluation of homework and exam sheets.