Dec 2024 – Feb 2025 · 2 mos · Remote

• Designed a real-time, machine learning-driven music system capable of recognizing ongoing musical beats and generating coherent continuations. This cutting-edge project leveraged state-of-the-art sequence prediction models—including LSTMs, Transformers, and diffusion-based models—to achieve remarkable accuracy in musical note prediction tasks.

Jan 2021 – Nov 2024 · 3 yrs 10 mos

• As a Machine Learning Expert, I collaborated with the UF Genetics Institute on a project aimed at developing a machine-driven pipeline to build a library of rAAV variants for facilitating drug delivery. The following milestones were achieved during this project:
• Developed a machine learning-guided pipeline for viral vector design, incorporating classification models for predicting rAAV assembly and generating a library of rAAV mutants.
• Trained a Transformer-based model on a dataset of approximately 21 million viral sequences, achieving 73% accuracy in capsid assembly prediction, representing an 8% improvement over benchmark deep learning models.
• Technologies: Python, Transformers, RNN, CNN, SVM, PyTorch, SciPy, NumPy, CUDA.

Aug 2018 – Nov 2024 · 6 yrs 3 mos

• As a Teaching Assistant in the Computer Science Department at the University of Florida, I have led discussion sessions, provided tutoring, held office hours, and developed and graded assignments in the following courses:
• Machine Learning Engineering (CAI4104/6108) – Spring 2024
• Advanced Machine Learning (CAP6617) – Spring 2024
• Machine Learning (CAP6610) – Spring 2023, Fall 2020, Fall 2018
• Applied Machine Learning (CAP6617) – Spring 2022, Spring 2021
• Data Structures & Algorithms (COP3530) – Summer 2022
• Math for Intelligent Systems (COT5615) – Fall 2022, Spring 2019, Fall 2018

Jan 2018 – Nov 2024 · 6 yrs 10 mos

• I am currently a Research Assistant in the Computer Science Department at the University of Florida, specializing in Computational Neuroscience. My research focuses on understanding how layers of spiking neurons influence the Volterra-based transformation from continuous-time signals to spike trains. In my most recent project, I have accomplished the following:
• Developed a novel framework for coding and decoding continuous-time signals using biologically plausible spiking neurons.
• Derived reconstruction bounds for the framework applied to a generalized class of Finite Rate of Innovation (FRI) signals.
• Validated the framework on 1,000 audio snippets, achieving a 𝟐𝟎 𝐝𝐁 𝐒𝐍𝐑 𝐚𝐭 𝟏/𝟓 of the 𝐍𝐲𝐪𝐮𝐢𝐬𝐭 𝐫𝐚𝐭𝐞. Optimized spiking neuron parameters using Stochastic Gradient Descent, leading to approximately 75% improvement in reconstruction accuracy.
• Presented my work at ICASSP 2023, Sigma Xi 2024, and a seminar at Dr. Joel Harley’s lab in the ECE Department at UF (November 2019).
• Technologies: Python, PyTorch, TensorFlow, NumPy, pandas, SNNTorch, Seaborn, Java, Git, Anaconda, PyCharm.

May 2017 – Dec 2017 · 7 mos · Gainesville, Florida Area

• As a Software Engineering Intern at Optym, I contributed to SkyWorks, a project aimed at facilitating the viewing, editing, and analysis of flight schedules for Amadeus, which serves 1.7 billion passengers annually. My responsibilities included:
• Designing data models to be used by the application and UI layers.
• Implementing various algorithms for performing queries and optimizations on flight schedule data.
• Executing performance and unit tests to ensure the efficiency and reliability of the models and algorithms.
• Technologies: C#, Visual Studio, JUnit, RESTful APIs, Design Patterns.

Aug 2013 – Aug 2016 · 3 yrs · Bangalore, India

• As Software Engineer (level 2) in the Enterprise Content Division at EMC corporation specializing in cloud and backend development, I have mainly made the following contributions:
• I have majorly worked as backend java developer, building the backend rest resources for a project called Capital Projects.
• Built backend REST resources for Capital Project Thermal, a cloud-based document solution on Documentum, supporting projects for hundreds of clients in the oil, gas, and construction industries, improving document handling by ≈30%.
• Refactored the monolithic REST solution into microservices, boosting scalability and cutting downtime by ≈40%.
• Technologies: Java, Spring, Documentum, Cloud, MongoDB, Docker, Kubernetes, JMeter, AngularJS.
• Awards: EMC ECD Hackathon Winner (2016), EMC Excellence Silver Award (2014).

May 2012 – Jul 2012 · 2 mos · Hyderabad Area, India

• As a 10 weeks' intern at Amazon Inc. (in Hyderabad, India) I was part of the Customer Services Team and mainly worked on a couple of projects:
• Developed a Java date and currency formatter utility module which would take any locale as parameter and generate the properly formatted corresponding date or currency strings, which in turn could be used in several payment related emails sent to the customers.
• Migrated the payment soft-decline emails to a Java based framework from an existing perl based framework and subsequently improved its content.
• Technologies: Java, Perl, JUnit, PowerMock, SOA, Perforce, Brazil Build, Apollo.