Krishna Kumar KA

Email: adhw4t@r.postjobfree.com Phone: 470-***-**** LinkedIn: Krishna Kanagal

Summary

Data scientist with three years of experience in developing and deploying scalable Machine Learning pipelines into production. Skilled in Machine Learning, Deep Learning and Statistics. Over ten years of previous experience in software development and client management. Proven track record of solving real-world problems using AI and Data Analytics.

Skills

Languages Python, R, SQL, COBOL, CICS

Packages TensorFlow, Keras, scikit-learn, Pandas, NumPy, OpenCV, Matplotlib Techniques Deep Learning, Natural Language Processing, Time series Analysis, Machine Learning Algorithms CNN, RNN, LSTM, word2vec, Linear Regression, SVM, PCA, Decision Trees, Logistic Regression, XGBoost

Database MySQL, DB2, Oracle, BigQuery

DevOps & Bigdata Git, Dockers, Kubernetes, Spark, Google Cloud Platform Tools Tableau, Informatica, Jira, Excel, Flask, Kubeflow Professional Experience

Farmwave LLC Senior Data Scientist May 2018 – Aug 2020 Farmwave is a company focused on solving problems in agriculture using Artificial Intelligence (AI). Counting kernels on corn cobs to calculate yield, identifying grain-loss during crop harvesting, and early identification of plant diseases are some of the problems tackled. My contributions are as follows:

• Built the first-ever plant pathology database after collecting images for four crops and assessed the quality of these images using a tool built from Python and OpenCV.

• Developed a Convolutional Neural Network (CNN) model to help early detection of crop diseases with a weighted precision of 92% and F1 score of 90% on evaluation images.

• Developed backend AI components for crop disease detection in the Farmwave mobile app and deployed it on google cloud using Docker and Kubernetes.

• Saved an average of $25 per acre by building a CNN model to estimate crop harvest loss with a precision and an F1 score of 87%.

• Achieved 8x faster inference and 85% memory reduction by optimizing the harvest loss model using post- training quantization techniques.

• Saved 8 minutes of manual work per corn cob by automating the kernel counting process using CNN with 96% accuracy.

• Contributed to building an IoT device used for harvest loss estimation and disease prediction and deployed the AI models on the devices using Docker and Flask RESTful.

• Created Tableau visualizations to showcase AI metrics such as precision, recall etc. to key stakeholders.

• Automated training and evaluation of AI models using Kubeflow pipelines.

• Increased sales of the IoT devices by exhibiting its capabilities to tractor manufacturing companies. Georgia State University Research Assistant Aug 2017 – Apr 2018 Westrock Project

• Read barcodes on shipments using OpenCV and extracted information using Tesseract OCR.

• Implemented Computer Vision algorithms to count the number of cardboards in the shipment stack. Cross-checked this number with the information from the barcodes to reduce mislabeling by 90%. SunTrust Bank

• Classified the quality of Anti-Money Laundering (AML) alerts of SunTrust bank with an accuracy of 81% using Random Forest model.

• Presented Tableau visualizations of exploratory data analysis and results to SunTrust executive team. Legal Analytics Project

• Achieved an accuracy of 76% in predicting the sentiment of the Securities and Exchange Commission

(SEC) documents using an LSTM neural network. Word embeddings were created using word2vec. Infosys Ltd. Technology Lead Jan 2015 - Aug 2017

• Established and lead a 19 member Data Warehouse and automation testing team for the American Express AML platform. The group generated an annual revenue of $1million.

• Interacted with the client business analyst team to understand and gather requirements for the Data Warehouse for AML platform.

• Lead the development and maintenance of the Data Warehouse built using Informatica to generate reports to assist AML compliance officers to take informed actions.

• Discovered repetitive tasks and guided the team to build automated scripts using HP UFT, which resulted in $50,000 in savings annually.

Dell Services Ltd. Senior Analyst Aug 2006 – Dec 2014

• Interacted with clients to understand the requirements of new projects and provide the status of on- going projects.

• Managed a 10-member offshore team from client location and maintained active communication between the team and business analysts to ensure timely delivery of critical projects.

• Migrated data of AmeriCredit automobile financing from legacy Databases like IBM DB2 and VSAM to Oracle Database using Informatica.

• Created complex mappings using Informatica to load data to a Data Warehouse and develop Business Intelligence reports for a major mortgage lender in the US.

• Identified performance issues in over 400 SQL queries and made changes accordingly to improve the query performance by 10%.

• Coordinated with business analysts and development team and translated business requirements to technical specifications.

• Fixed an inefficient and time-consuming batch process that resulted in a savings of £200,000 in penalty.

• Implemented COBOL code according to design specifications for a credit rating agency, conducted testing and deployed to production.

Education

Master of Science in Analytics Aug 2017 - Dec 2018 Georgia State University, Atlanta, GA GPA - 3.8

Bachelor of Engineering, Mechanical Engineering Sep 2002 - Jun 2006 National Institute of Engineering, Mysore, India GPA - 3.5 Certifications

Coursera: Machine Learning: Regression Sep 2020

Coursera: NLP with Classification and Vector Spaces Oct 2020 Coursera: Practical Time series analysis Nov 2020