Ashwin Kumar U. Email & Phone Number

San Francisco, Ca, Us

New York, Us

LLMs, Vector DBs, Full stack engineering

New York, Ny, Us

• Leadership: Served as the CTO and implemented a technology roadmap aligned with business adjectives while leading a team of 8 software engineers, resulting in a 70% increase in operational efficiency.• Product Development: Led the development of knowledge.com, and went from whiteboard to beta stage within 4 months. Re-architectured the application using Typescript & MVC design, reducing crashes by up to 80%.

Brooklyn, New York, Us

• Computer Vision: Responsible for developing target tracking algorithms and optimization using TensorRT and Pytorch. Developed TinyML models to be deployed in Nvidia Jetson boards.• Controls: Duties include implementing algorithms through ROS, including communication from CAN bus.

New York, Ny, Us

Lead Engineer for DemandCapture's upcoming SaaS platform - knowledge.com. Leveraging data to provide targeted marketing using Machine Learning and Deep Learning Techniques.• Leadership: Led the backend and frontend development of the SaaS platform while managing an overseas team, and accomplished implementing 35+ features. Communicating and facilitating technological partnerships with Executive and VP-level employees in publicly traded companies.• Proprietary Deep Learning Algorithm: Developed a proprietary model that takes in Categorical, Numerical, and time-series data and beats the results of the SoTA model by 10%.• Bombora Partnership: (https://bombora.com/integration/demandcapture/) Spearheaded the partnership engineering with Bombora and automated the movement of about 2.5 billion rows of data per month to Google BigQuery. Wrote SQL scripts that sped up the movement process from 30+ hours to just 3 hours.• ZoomInfo Partnership: Led the partnership engineering efforts with ZoomInfo, managing and distributing work with 3 developers in another timezone.• Open-AI models: Finetuned OpenAI models such as GPT 3.5-turbo using internal data records.• Recommendation Engine: Designed an embedding engine that creates embedding representations for prospects and sales team and gives the sales team the most relevant prospects to talk to.• Backend Engineering: Developed a gateway server that interacted with various other microservices and reduced the client load speed by more than 40% by leveraging technologies such as Redis, load balancing, and job queues.• Security Engineering: Implemented various security measures such as data encryption on the Database, key rotation, password encryption salt, and Hardware Security Module, and proposed a Data Breach Handling Protocol in compliance with the Federal Trade Commission (FTC) recommended steps.

New York, New York, Us

TA under Prof. Hesam Oveys, teaching topics on Data Science, Statistics and Machine Learning.

New York, New York, Us

Roles Include:· Hold recitation during the scheduled time and help students understand and practice the material taught during the lecture.· Holding office hours.· Prepare and grade materials such as quizzes and worksheets.· Help proctor and grade midterm and final exams.· Communicate and coordinate with the course instructor.

Chennai, Tamil Nadu, In

· Deep Learning based Predictive pre-fetch: Predicts and prefetches the next page based on the user activity, reducing the load speeds by 35% and analyzed the feature with V8 stack traces and optimization phases to identify bottlenecks.· Performance log parser: Reverse-engineered Google chrome to understand how performance and function trace logs are parsed and devised a multi-threaded parser that reduced parsing time from 25 mins to about 1.5 minutes.

Chennai, Tamil Nadu, In

· V8 compiler analysis: Assessed Zoho CRM performance, and inspected the function compilation, optimization, and performance on the V8 compiler. Advised steps to streamline the functions to improve API performance and minimize redundancy.· CPU throttling analysis: Developed a bash script to throttle CPU performance and analyzed the performance of Zoho CRM to detect about 150+ CPU heavy functions and persuaded for coding into lighter functions to minimize client load.

· Ubuzz: Headed the development of a blog application with markdown support on Android. Designed an audience monitoring dashboard, and an editorial dashboard that allowed the 20+ internal editors in the company to publish articles seamlessly.· PlacementPrep: Headed the development of a mobile-based MOOC platform entirely. Developed audience monitoring and course management dashboards for the Edutech startup. Initially opened to 15+ universities, and then launched publicly.

· Developed E-Commerce (MOOC) android applications and supporting dashboards.· Designing and prototyping User Interfaces in Adobe XD.

Chennai, Tamil Nadu, In

Designed and implemented an intelligent predictive-prefetch feature using Deep learning and other specially devised methodologies.

In

Internship on cloud computing

Master Of Science - Ms, Computer Science

Bachelor Of Engineering - Be, Computer Science