Sep 2021 – Nov 2024 · 3 yrs 2 mos · Munich, Bavaria, Germany · Hybrid

• Architected enterprise-scale ML platform using Kubeflow on Kubernetes, enabling 10+ teams to implement CD4ML pipelines with automated model deployment
• Orchestrating successful delivery of 3 production ML use cases, driving initiatives to achieve 20% reduction in use case-specific operational overhead by 2024
• Pioneered Generative AI capabilities by implementing open-source model hosting and RAG architecture using in-memory vector databases, empowering teams with production-ready LLM solutions
• Tools & Technologies:
• Kubernetes, Kubeflow, KServe, Terraform, Python, AWS Services (IAM, Glue, S3, EMR on EKS, SageMaker, Athena), Generative AI (LLMs, Multimodal Images), RAG architectures

Aug 2019 – Aug 2021 · 2 yrs

• Engineered scalable ETL pipeline processing 150+ GB daily data across multiple sources, unifying user interactions, impressions, and revenue data
• Unified siloed data systems by implementing modern data mesh architecture using Snowflake and DBT, enabling cross-functional content analytics
• Drove 12% increase in annual revenue through automated content performance analysis, identifying high-value article types and engagement patterns
• Tools & Technologies:
• AWS Athena, AWS Glue, AWS EMR, MWAA(Managed Airflow), Apache Spark, Snowflake, DBT, Python, Terraform, Gitlab CI

Oct 2018 – Jul 2019 · 9 mos

• Architected and implemented serverless AWS infrastructure using CloudFormation, processing 10M+ daily user interactions and 100K+ news articles for recommendation engine
• Optimized cloud costs by 35% through event-driven processing and intelligent S3 lifecycle policies
• Designed scalable data pipeline using AWS services (Lambda, Kinesis, DynamoDB, Aurora Serverless, Redis) enabling data scientists to reduce model training time by 40%
• Tools & Technologies:
• AWS cloud, Lambda, Kinesis, Apigateway, Elasticache, Dynamodb, Firehose, EMR, Python3, Scala, YAML

Feb 2017 – Sep 2018 · 1 yr 7 mos · Neu-Isenburg

• Architected and delivered a scalable data pipeline using Apache Spark and Kafka, replacing legacy monolithic system, reducing processing time by 35%
• Built mission-critical ETL workflows for data enrichment and complex business logic implementation, processing huge amount of daily transaction data
• Integrated COTS solutions with the Hadoop ecosystem, providing data to these systems in near-real time
• Tools & Technologies:
• Java, Apache Spark, Kafka, Hadoop (HDFS, YARN), Solr, HBase, Phoenix, Nifi, Flink, Oozie, HDP 2.5, CDH 5.13, Azure Cloud

Oct 2016 – Sep 2018 · 1 yr 11 mos · Sankt Augustin

• Working on the Big Data Technology stack which includes Spark, Hive, Impala, Sqoop, Kakfa and Flume.
• Setting up of virtual hadoop/CDH cluster using Vagrant and Ansible.

Mar 2016 – Sep 2016 · 6 mos · Berlin Area, Germany

• Working on distributing big datasets into small manageable chunks using Python and later, would be testing them on scalable systems like Flink and Spark to run machine learning tasks.

Oct 2012 – Sep 2014 · 1 yr 11 mos · Gurgaon, India

• Underwent 3 months of extensive training in C, Data Structures, Unix, and, System Programming. Also basics of Telecom and Networking.
• Worked for Multinational data networking and telecommunications equipment company, building and maintaining features used in PCU(Packet Control Unit) for 2G networks

Jun 2010 – Jun 2010 · 0 mo · The Air Force School, Subroto Park

• Our Task was to study about the structure of .bat files, So as to infect them and then device a way to restore the original file.