Prathik Vijaykumar

YahooMountain View, California

Data Engineer InternMay 2023 - August 2023

Yahoo is a global media and tech company providing various services, including a web portal, search engine, news, email, and advertising services. During the internship, I was part of data plaforms team tasked with user-profiling and I worked on building core datasets, enabling data-driven decision-making on Yahoo's product improvements.

• Architected AWS QuickSight dashboards, improving cloud security & operational efficiency, improved cost utilization & data visibility.
• Developed an AWS SNS-based system for real-time data anomaly alertsin S3 buckets, boosting data integrity and quality check by 25%.
• Automated data ingestion & optimized API queries to retrieve user data from Gemini, achieved 96% accuracy
• Developed a BERT-based NLP model chatbot, reducing internal data query search time by 30%.
• Skills: AWS QuickSight, Terraform, Airflow, DynamoDB, Python, Streamlit, BERT, SNS

Texas A&MCollege Station, Texas

Machine and Cloud researcherSeptember 2022 - Present

The Learning and Emerging Networked Systems (LENS) at Texas A&M aims to conduct analytical and experimental research in machine learning and reinforcement learning with networked systems applications.For 1.5 years, I worked as researcher under the resource optimization department. We are Trying to disrupt the field of orchestration and autoscaling system for allocating resources to microservice applications. Using zeroth gradient approaches and reinforcement learning methods to achieve these objectives.

• Optimized data processing pipelines for a social media microservice application, achieving a 12% reduction in latency.
• Utilized Docker for efficient containerization and integrated real-time data monitoring with Jaeger. Used reinforcement learning and policy gradient strategies for effective resource management, enhancing system scalability to exceed industry standards by 14%.
• Extended microservices with Spring REST APIs, using Docker for containerization, Kubernetes for orchestration, and Nginx/Apache Tomcat for load balancing and secure token-based authentication.
• Integrated real-time monitoring and ETL pipelines, reducing data processing time by 40%
• Skills: Docker, Kubernetes, Spring, Nginx, Apache Tomcat, Spark, Telemetry, Grafana, Jaeger

Texas A&MCollege Station, Texas

Machine and Data EngineerSeptember 2022 - Present

• Deployed machine learning models for plant stress detection, achieving 95% IOU with YOLOv8.
• Architected full-stack web service for heat risk assessment, integrated databases, containerized and deployed it
• Integrated MLOps pipelines for automation and scalability.
• Skills: PyTorch, Django, Flask Docker, AWS SageMaker, NoSQL(MongoDB, DynamoDB), MySQL, Airflow, ECS.
• Skills: Docker, Kubernetes, Spring, Nginx, Apache Tomcat, Spark, Telemetry, Grafana, Jaeger

Redbus (Bus & Train Tickeing Application)Bengaluru, India

Data EngineerJuly 2020 - July 2022

RedBus is an Multi-National online bus ticketing platform that facilitates the booking of bus tickets and other related services across multiple cities and operators. I worked for two years as a Data Engineer. I was part of the data lake team where i contributed in creation of data lake for unified view, developed ETL pipelines for inventory management, sales, payments & Marketing (CRM channels). Additionally developed dashboards & websites to drive the adoption of premium services offered to operators.

• Developed and scaled 40+ ETL pipelines with AWS, Apache (NiFi, Kafka, Spark, Storm, Hadoop), processing 10M events daily. Optimized batch processing with Spark (8M+ records), saving 2 hours runtime.
• Migrated legacy systems to AWS, creating a data lake with Redshift (15% retrieval time reduction, 25% storage cost savings). Warehousing (Redshift). Developed dashboards tailored to business needs. Worked with crossfunctional team from India and Singapore.
• Engineered a Marketing and Sales Data Analysis Platform, integrating hot and cold data storage (PostgreSQL, AWS S3) and dynamic dashboards (Tableau, Redash), leading to a 2.5% increase in customer acquisition and an 8% rise in quarterly transactions. Under this initiative, built a real-time user interaction analytics pipeline (Kinesis, DynamoDB) processing 400K mobile app events hourly, delivering user engagement insights.
• Built real-time pipelines for user interaction data on Android and payment analytics, processing 400K events hourly and reducing failure notification time by 80%, using Kinesis, Lambda, DynamoDB, S3, SNS, Kafka, and Spark-stream.
• Architected data-driven rating system for 3,500 bus operators across India and Singapore, implementing ETL pipeline with cold/hot storage in Redshift/DynamoDB. Developed 3NF-compliant models, Tableau dashboards, and Django REST API-powered web app. Resulted in 20% increase in Premium service adoption among operators
• Deployed surge pricing & anomaly detection models using AWS SageMaker (CI/CD with Kubeflow, Jenkins, GitHub Actions). Achieved 25% faster model updates, leading to $1M revenue increase & 40% fraud detection improvement.

InternJan 2020 - June 2020

During my internship i worked on wide array of IoT,NLP, computer-vision based projects as i was part of sole R&D team.

• Developed an IoT - computer vision based solution for real-time bus footfall tracking, resulting in a 17% reduction in pilferage.
• Created onboard diagnostics system to capture & display bus performance. Reduced daily fuel consumption by 500 gallons.
• Created an Alexa application & streamlined ticketing for voice-command bus ticket bookings, serving 20000 customers.

Electronics Cooperation of India LimitedHyderabad, India

InternJune 2018 - August 2018

Electronics Corporation of India Limited (ECIL) is a Government of India enterprise that develops and manufactures electronic products for the defense, nuclear, and industrial sectors.

• Designed a dashboard using grafana and JSON-API to visualize the load characteristics of a local electric grid.