United States Machine Learning

Kowshika Reddy

716-***-**** *.******.*****@*****.*** LinkedIn Profile GitHub

Education

University at Buffalo, The State University of New York Aug 2023 – Dec 2024 Master of Science in Computer Science (Data Science), CGPA: 3.9 Buffalo, NY Experience

University at Buffalo Oct 2024 – Dec 2024

Graduate Research Assistant Buffalo, United States

• Engineered a machine learning model using bagging and boosting algorithms (Random Forest, XGBoost) to categorize crime types in Los Angeles, addressing the imbalanced dataset and improving prediction accuracy by 25%.

• Applied hotspot clustering techniques like DBSCAN and HDBSCAN to identify high-risk crime areas, enabling targeted interventions and better crime prevention strategies based on spatial analysis.

• Utilized LDA for uncovering latent topics in crime descriptions and clustered these topics using BIRCH, Mini Batch KMeans, and OPTICS, combined LSTM models to analyze and forecast crime fluctuations, improving future crime prediction. One Park Financial May 2024 – Oct 2024

ML Software Engineer Intern Texas, United States

• Developed a hybrid content-based recommendation microservice (TF-IDF + LightGBM) using Flask, leveraging RabbitMQ for real-time loan offer delivery, boosting message throughput and reducing latency by 20%.

• Optimized 50+ SQL functions in PostgreSQL for customer segmentation, integrating React dashboards with the Tableau JavaScript API, improving dashboard load times by 30% for accurate decision-making.

• Maintained data pipeline uptime of 99.8%, while ingesting streaming and daily transactional data from API’s, using Spark, Snow Flake and Python, enabled targeted promotional offers SMS via Twilio API (Restful API).

• Achieved a 5% increase in ROI for financial product promotions using A/B testing in Google Optimize, integrated with Google Analytics, to optimize channel and messaging strategies for targeted customer segments. Contriver Ltd Aug 2020 – Apr 2023

Data Engineer Bengaluru, India

• Migrated on-premises data storage from Microsoft SQL Server to AWS Redshift by S3, enabling scalable storage for 5M+ oncology patient records, cutting storage costs by 30%, and enhancing backup reliability.

• Orchestrated the construction of data ingestion framework through AWS Glue to facilitate smooth transitions of large volumes of data from different sources directly into Amazon Redshift, for data warehousing and analytics.

• Optimized SQL queries in Redshift through CTEs, indexing, query execution plans, distribution keys, sort keys, and Parquet format, minimizing execution time from 10s to 7s for faster data retrieval for data analysis.

• Deployed a real-time data pipeline leveraging Kafka and PySpark to process semi-structured data from 30+ sources, automating workflows with Apache Airflow, increasing records processed per minute by 40%.

• Automated CI/CD pipelines operating Jenkins to streamlining the deployment of Docker images for external business clients, diminishing deployment time by 30% and ensuring reliable delivery of data pipelines.

• Designed Power BI dashboards using Power Query, and DAX to identify data discrepancies, reducing reporting time by 20%, and automated data refresh with scheduling tools.

Projects

Smart Meal Planner: AI-Powered Recipe Generation and Ingredient Optimization Link

• Developed a smart meal suggestion system using YOLO (v7, v9) for ingredient detection and ChefTransformerT5 (Hugging Face) for recipe generation, and improving ingredient recommendation accuracy by 25%, displaying nutritional values via API.

• Integrated BART embeddings, Grounding DINO, and NER for ingredient identification, reducing recipe generation latency by 15%, enhancing efficiency, and enabling personalized meal suggestions that complement users pantry items. AI-Driven Stock Market Prediction and Analysis System Link

• Built an RNN-based stock prediction model using historical data fetched from the Yahoo Finance API, leveraged Long Short- Term Memory to capture long-term dependencies in stock price movements, improving prediction accuracy by 20%.

• Deployed the stock prediction model on Databricks, used Tableau dashboards for real-time visualization of stock trends and performance metrics (RMSE, MAPE), enhancing data-driven decision-making and improving strategic decisions by 25% Skills

Programming Languages: Python, R, SQL, Scala, Java, DAX. Databases: MySQL, PostgreSQL, MongoDB (No SQL).

Big Data Tools: Apache Hadoop, Apache Spark, Apache Kafka, Hive, Terraform, Informatica. Packages & Frame Works: TensorFlow, NumPy, Pandas, Open CV, Scikit – learn, Keras, Pytorch, Fast API, Rest API, GraphQL. Cloud Platforms: AWS S3, IAM, EC2, EKS, Kinesis, RedShift, Sage Maker, Glue, DMS, DML, Data Bricks, Snow Flake. Developer Tools: Power BI, Excel, Tableau, Looker, CI/CD, MLOps, Docker, Kubernetes, Git, Bitbucket, Jupyter Note Books. Other: Data Analysis, Data Mining, Data Processing, Data Modeling, Data Visualization, Predictive Modeling, Statistical modeling, CUDA, NLP, ANN, CNN, Agile Methodology, Transformers, OOPS, SDLC, Code Review, Version Control, Unit testing, Debugging. Certifications: Meta Back- End Developer AWS Certified Data Engineer – Associate Tableau Analyst.

Contact this candidate