I'm Ramkumar, welcome to my portfolio! My fascination for the applications of Artificial Intelligence in product and
business use cases pushed me to pursue my master's in Artificial Intelligence at San Jose State University.
Since AI is a diverse field with different focus areas, my goal was to widen my skillset covering most sub-fields,
such as Computer Vision, Natural Language Processing, Data Science, etc. During the Summer of 2022, I secured an internship at Veritas Technologies,
where I worked on accelerating business with Machine Learning. With this stint, I gained experience applying Data Science to business use cases and
converting a project from an experimental to a development stage by attracting internal stakeholders. Currently, I'm working on two projects in
NLP: A Semantic Search algorithm for Question-Answering on Personal Files and an Application Tracking Assistant.
- Identify business use cases with data science and present them to stakeholders.
- Acquire a raw dataset, explore, visualize, identify the appropriate state-of-the-art model, train and deploy the ML model in production.
- Work with the text, image, and audio data. Build data pipelines for continuous training and inference.
Resume Preview
Technical Skills
Programming Languages:
Python, SQL, C++, R Programming, C, JavaScript, Java, MatLab.
Tools/Framework:
TensorFlow, PyTorch, NumPy, Pandas, Sci-kit Learn, Keras, OpenCV, Tableau, PowerBI, Git, Jupyter, Flask, Excel, RASA, Apache Spark, Microsoft Azure, Hadoop, Linux, Matlab, Azure, AWS.
Education
San Jose State University, San Jose, California, USA
| Master of Science in Artificial Intelligence | 3.95 CGPA |
SRM Institute of Science Technology, Chennai, India
| Bachelor of Technology in Information Technology | 8.3 CGPA |
Work Experience
San Jose State University | Teaching Assistant | Sep 2022 - Dec 2022 |
- Guide undergraduate-level students in applications of Artificial Intelligence using Python.
- Assist students in applying ML libraries during lab sessions and office hours.
Veritas Technologies LLC | Business Intelligence Analyst Intern | May 2022 - Aug 2022 |
- Built Machine Learning models and recommender systems for Veritas’s products to identify cross-sell opportunities
and optimize pricing strategy with cross-functional teams by modeling customer behavior.
- Constructed end-end ETL and ML pipelines in Azure with data from different sources using SQL and PySpark and
reduced the cost of spark instances by optimizing the model’s runtime by 50%.
- Designed interactive Power BI dashboard/API with python scripts to visualize and identify sales opportunities for
business. This enabled teams to interpret product associations and find sales opportunities faster.
- Winner of Intern Project Showcase competition, where Veritas’ interns in the US showcase their projects.
Rise Lab, Indian Institute of Technology Madras | Research Intern | Dec 2019 - March 2020 |
- Implemented an exploratory analysis on gate-level netlists of different TrustHub benchmarks (Hardware Designs) by collecting data through parsing and simulations.
- Built Machine Learning models to detect Trojan nets from Normal nets during the IC design phase with 90% TPR to enhance security.
Allsec Technologies | Data Analyst Intern | June 2018 |
- Implement exploratory analysis on the employee survey data using R Studios to support business teams.
- Built regression and random forest classifier in R to predict employee performance metrics.
Work Experience
SEMANTIC Q/A SEARCH (Ongoing Masters Project) | Aug 2022 - Present |
- A local NLP search application to retrieve files and answer user queries semantically based on information extracted from documents using a BERT-based model and knowledge graphs.
HYBRID MUSIC RECOMMENDER SYSTEM | Sep 2022 - Dec 2022 |
- Built a hybrid system to recommend songs with Sentence Transformers and KDTree to identify similar songs based on the lyrics, an SVD Matrix Factorization model to predict the “play count” of songs per user, and content-based collaborative filtering models to rank songs based on the metadata.
- The hybrid system predicted the play count of a song with an MAE of 0.4 and suggested songs with 43% accuracy on a custom evaluation method on test users.
- Designed a UI to recommend songs using streamlit.io API.
MASK2FACE USING U-NET GENERATIVE ADVERSARIAL NETWORKS | Jan 2022 - May 2022 |
- Trained U-Net Deep Learning models using Tensorflow to generate facial features for masked human faces.
- Synthesized unmasked RGB and Grayscale faces for input images of masked human faces with an error rate of 0.3.
- Built an application to identify human faces with masks and synthesize facial features to identify the real person.
CRIME SAGE - LOS ANGELES CRIME PREDICTOR | Sep 2021 - Dec 2021 |
- Developed a model that suggests the most probable crimes in a location with a travel safety index.
- Obtained an accuracy of 80% in predicting the top 3 probable crimes using Sequential Neural Networks using Tensorflow based on personalized features of the user.
PERSONAL FARM ASSISTANT - A CHATBOT | Aug 2021 - Dec 2021 |
- Built a chatbot using Rasa 3 to help agriculturists and farmers in managing multiple farms efficiently.
- The bot monitors crop growth and climatic conditions; reads messages from workers and controls automated functions using contextual conversation using NLP.
Accomplishments
PUBLICATION - (IEEE SPICES 2022)
“Deep Convolution Neural Network Analysis For Crop Growth Stages Prediction”, Communicated to IEEE International Conference on Signal Processing, Informatics, Communication and Energy Systems 2022
PATENT - A SYSTEM FOR CROP GROWTH PREDICTION AND METHOD THEREOF
Indian Patent Office - 202041052748A
Recent Work
NLP BERT based Lyrics Similarity Model + SVD Play Count Regression Model.
A Text analysis on the 'Consumer Complaint Database' dataset to build a classification model with NLP.
Cricket Analysis using Web Scraping and Data Science.
Analysing the performance of the player using Python.
Get In Touch
I'm actively looking for full time Data Science, Machine Learning and Business Intelligence roles. Reach out if you have a suitable role.
