Software Engineer Machine Learning

UMESH KUMAR GATTEM

Artificial Intelligence Engineer Machine Learning Software Engineer Associate Instructor of Statistics Tracy, CA, 95391 Mobile : +1-812-***-**** e-mail: *****.******@*****.*** Portfolio: https://umesh-gattem.github.io/ Linkedin: https://www.linkedin.com/in/umesh-kumar-gattem-8b9142a8/

● An AI engineer having 5 years of experience, proficient in designing tools for visually constructing data transformation recipes, DL/ML models, and end-to-end pipelines, while consistently taking ownership and responsibility for projects. EDUCATION

Master of Science in Data Science Indiana University of Bloomington (GPA: 3.8/4.0) Aug 2021 - May 2023 Related Courses: Deep Learning, Machine Learning, Statistics, Algorithms, Computer Vision, NLP, Big Data Bachelor of Technology in Computer Science and Engineering ANITS (GPA: 8.05/10.0) Aug 2012 -Apr 2016 TECHNICAL SKILLS

Languages - Python, C, C++, Java, HTML, Javascript, R Frameworks - TensorFlow, PyTorch, Keras, Sci-kit, Pandas, Numpy, Flask, OpenCV, Pyspark, FastAPI, Uvicorn Machine Learning - PCA, T-SNE, TF Hub, Neural Networks, Clustering, Transfer Learning, Inferring models Databases and Big Data - MySQL, Postgres, MongoDB, SQLite, Spark, Hadoop, Kafka, S3, Horovod ML Ops: Docker, Kubernetes, GCP, AWS, Distributed Training, Tensorboard, Weights, and Bias Project Management and Tools - Git, JIRA, Confluence, Slack, Pycharm, IntelliJ, Jupyter Notebook PROFESSIONAL EXPERIENCE

Associate Instructor - INDIANA UNIVERSITY BLOOMINGTON, Indiana Aug 2022 - May 2023

● Aided in Introduction to Statistics and Applied Statistical Computing courses, supervising graduate and Ph.D. students by resolving doubts, conducting office hours, creating assignments, grading, and providing feedback and solutions. Machine Learning Engineer - RAZORTHINK TECHNOLOGIES, Bangalore, India Jan 2018 - Jul 2021

● Developed RZTDL Library which is a patented deep learning framework built on top of different backends of Tensorflow and PyTorch supporting all Deep Learning operations like CNN, RNN, LSTM, and Attention.

● An integral contributor to the 10-people engineering team guided in different aspects including model design, training, loading, inferring, and transfer learning of models from the in-house DL model store, called MLC (Model Lifecycle).

● Facilitated the development and enhancement of user-friendly DL blocks, simplifying the process of constructing complex models and empowering end users to leverage deep learning techniques more efficiently in the projects.

● Designed a multi-stage CNN model for analyzing multidimensional time-series data, resulting in a GINI score of 72 for predicting customer propensity to buy insurance for esteemed banking clients.

● Implemented LSTM network to develop a churn model by conducting a comprehensive analysis of demographics and skewed transactional data, resulting in a GINI score of 68 for one of the largest telecommunications clients. Software Engineer - RAZORTHINK TECHNOLOGIES, Bangalore, India Jul 2016 - Jan 2018

● Created distinct Python web applications tailored to the product requirements, seamlessly integrating front-end and back-end APIs with the utilization of Flask API, Uvicorn, and FastAPI.

● Led a team of 4 skilled professionals in developing a Python SDK, which facilitated the generation of blueprints, automated the process of generating code, and created APIs for seamless integration with web applications.

● Built diverse data transformation modules to capture and flow substantial datasets through different blocks like JOIN, and Group By and used Kafka for data streaming which accelerated data flow by over 100% of traditional streaming.

● Supervised 5 member team, implementing Python libraries with production-level code, ensuring the readiness for deployment every weekend by pushing the code to Staging and QA environments using nightly versions. PERSONAL PROJECTS:

Facial Recognition on Google Photos: Github Poster Feb 2023 - Present

● Demonstrated an end-to-end model for facial detection by leveraging Google Cloud APIs, YOLO object detection, and MobileNet architectures, all integrated into a proprietary Google Photos system and achieved an accuracy of 92%. Ask me About GANs, Encoders, Transformers, Attention, GPT-3, BERT, and End-To-End pipelines.

Contact this candidate