Dhvani Kothari

I'm

About

Bridging the gap between cutting-edge generative AI and real-world applications—crafting scalable ML solutions that turn creative potential into production reality.

Phone: +1 716-936-4705

City: California
Email: dhvanikothari03@gmail.com

Interests

Gen AI

Data Engineering

Machine Learning

Cloud Computing

Agentic Systems

Visualization

Algorithms

Natural Language Processing

Education

MS in Computer Science

August 2023 - Present

Relevant Coursework

Data Intensive Computing
Algorithms for Modern Computing Systems
Machine Learning
Deep Learning

BE in Computer Technology

June 2015 - May 2019

Relevant Coursework

Advanced Data Structurs
Software Engineering
Design and Analysis of Algorithms
Object Oriented Programming
Operating Systems

Experience

Brainchip Inc.

February 2024 - Present

ML Solutions Architect

Built RAG pipeline with FAISS integrated into Gradio, cutting retrieval latency to 500ms, boosting contextual Q&A accuracy by 35%.
Developed FastAPI microservices to orchestrate ASR, LLM, and TTS APIs for real-time multimodal voice interaction on edge devices.
Implemented CenterNet-based palm detection model for an interactive Flappy Bird demo, enabling real-time gesture control with >90% detection accuracy.

Brainchip Inc.

June 2023 - Present

Solutions Architect Intern

Empowered technical sales teams with FastAPI web-based GUI tool, driving 3x efficiency and customer engagement driving upsell.
Streamlined integrated circuit scaling estimation process, reducing 30% processing time, elevating data accuracy by 15%.
Automated unit testing with Pytest and Drone, reducing time-to-market by 2 weeks, elevating user-centric testing productivity.

Walmart Global Tech India

May 2021 - July 2022

Data Engineer

Developed time-series ensemble forecast solution for 599 Sam’s Clubs, saving ∼$1.2 million in annual operation planning costs.
Implemented real-time monitoring for critical model performance metrics, reducing model drift and data quality issues by 20%.
Led a team of 6 to develop ETL mappings and data modeling on GCP, empowering data-driven insights for business derivatives.
Revamped PySpark pipelines with partitioning and query optimization, giving 20x speed improvement, saving $500k annually.
Spearheaded migration of on-premises data pipeline handling 1.5 TB data to GCP using GCS Bucket, Cloud Dataflow/Dataproc, Pub/Sub, Bigquery, enabling 3x faster data processing.

Capgemini

June 2019 - April 2021

Big Data Developer

Optimized Spark applications, achieving 30% faster processing through batch intervals, parallelism, and query enhancements.
Pioneered holistic automation using Azure Logic Functions and Azure Data Factory, delivering 100 man-hours in quarterly savings.
Designed a comprehensive data lake spanning 11+ manufacturing sites, enabling seamless data access for over ∼150 data scientists.
Engineered generic data recovery library, resulting in a 50% reduction in storage & memory costs and annual savings exceeding $100K.
Conducted workshops for business teams, enabling them to leverage data insights, improving the data literacy of the organization by 25%.