Suraj Kumar

Experience

• 

  IIIT Hyderabad

  May 2018 - Apr 2020

  Data Engineer

  - Devised LSTM architecture for word-level and character-level language modeling, handling multiple scales of vocabulary, and incorporated strategies with 3+ Ph.D. scholars to reach state-of-the-art performance.- Conducted experiments over richly agglutinative Indian Languages, the amalgamation of word embedding and syllable embedding with LSTM has shown 40% better performance than existing traditional methods.- Designed Azure data platform merging relational/NoSQL with Azure Scheduler, enhancing integrity by 40%, ensuring 99.9% uptime Show less

• 

  Google

  Jul 2019 - Apr 2020

  • Explore ML Facilitator

    Aug 2019 - Apr 2020

  • Developer Student Club

    Jul 2019 - Aug 2019
• 

  LTI - Larsen & Toubro Infotech

  Mar 2020 - Aug 2022

  Senior Data Engineer

  • Engineered an ETL-focused Airflow data pipeline, channeling vendor file system through AWS SFTP to the S3 landing layer. Leveraging PySpark for advanced transformations, achieved a 20% speed rise in data flow to the S3 curated layer• Overhauled the approach to transient EMR clusters with Airflow, initiating them for data processing and terminating post-task. Combining to Snowflake external table integrations, led to 18% reduction in operational costs, supporting efficient Product Lifecycle Management.• Collaborated with 4+ clients to migrate their data from local premise, AWS-Redshift to Snowflake, enhanced auto-maintenance by 40% Show less

• 

  University of Maryland Baltimore County

  Feb 2023 - May 2023

  Senior Research Assistant, MLLI dept. at B.E.A.R.D. Laboratory

  -Optimized text search efficiency by 30% with a customized web application featuring categorized difficulty levels.- Developed an advanced NLP software for Spanish texts that evaluates document readability, generates word similarities, and categorizes modules based on readability levels. Achieved 90% accuracy in assessing document readability.- Fabricated advanced ML models to classify readability levels with 90% accuracy using diverse feature combinations.

• 

  Webolinx

  May 2023 - Aug 2023

  Data Engineer | Freelance

  • Teamed up with data scientist, analysts, content managers, stakeholders to develop a CI/CD pipeline (GCP), utilizing business analytics insights for data-driven insights. Leveraged expertise in SAP, Teamcenter, SSIS, MS-SQL for process improvement and increased efficiency.• Automated data migration to AWS, moving 50M+ records, which optimized analytics and improved cross-border transit times by 15%

• 

  Oculi

  May 2023 - Dec 2023

  Senior Research Assistant

  • Designed data pipelines for CV, used CI/CD and Docker for consistent testing and deployment conditions, enhancing reliability of AI models• Optimized image recognition workflows: ingestion frameworks with Kafka, Spark in docker for latency and accuracy via Gitlab integration

• 

  FedEx Dataworks

  Jan 2024 - now

  Data Intern

  • Implemented a workflow using Apache Airflow (XCom), Apache NiFi and AWS (EC2, S3, CLI), streamlining extraction from 2+ million records. Used Snowflake for warehousing, cutting ingestion time by 15%, enhancing efficiency and cost savings by 30%• Collaborated with a data architect to design and implement an efficient data schema, including experience with SSIS for legacy data source integration enhancing data accessibility and performance. Leveraged CDC with SCD Type-2, gave 50% reduction in data inconsistencies.• Employed Agile methodologies for daily scrums and bi-weekly sprints to refine ETL pipelines with AWS Step Functions, which streamlined data deliverability and reduced project delivery timelines by 30%, ensuring timely and effective solution deployment. Show less

• 

  Bwtech@UMBC Research and Technology Park

  Feb 2024 - now

  Software Advisor II - Data Engineer

  • Architected a real-time data pipeline for high volume Ladybug sensor data, utilizing GIT for SCM, Kubernetes for deployment, Jenkins for CI/CD, and Databricks with Python, Kafka, and Spark for analytics. Improved surgical precision and patient safety by 45%• Deployed Azure ETL pipelines, transforming unstructured sensor data into physical Data Model (in Data Lakes), for efficient data analysis• Reported directly to CTO: Engineered a real-time surgical dashboard using Azure Synapse, enabling doctors to visually monitor tissue ablation during procedures. Utilized Omelek simulations to optimize surgical dashboard visualizations for real-time decision support. Show less