Sparsh Gupta Email & Phone Number

New York, NY, US

Mountain View, California, US

Expanding my role into driving initiatives end-to-end within the org.- Leading a 5-member team including engineers and data scientists to develop and deploy an LLM based model to extract data from customer tax documents. The team is working on novel LLM training strategies to optimize model robustness and accuracy, optimizing LLM hosting strategies to meet.

Mountain View, California, US

Most of my work is focused on NLP, with other modalities of data wherever possible - clickstream, tabular, and vision. The job responsibilities include collaborating with business partners to identify the ML application areas, obtaining and processing the data, ML modeling, and deploying and A/B testing the models. I have applied for 5 patents across all.

Mountain View, California, US

- Worked on a text-based classifier using LSTM and customer features to identify intent from customer questions.- Worked on a POC which was a suite of open-source models for customer chat analysis - sentiment analysis, topic modeling, and text summarization.

Mountain View, California, US

- Worked on question answer ranking models and FAQ search models using deep learning techniques like LSTM, attention mechanism and adversarial learning.- Worked on adversarial domain adaptation for question deduplication and recognizing textual entailment tasks. Published a paper on the same.

La Jolla, CA, US

I am a TA for two courses this quarter - CSE 190 Deep Learning, and CSE 276C Mathematics for Robotics. My task in the Deep Learning course is to prepare some of the assignments and evaluate the code and report submissions, in addition to holding office hours and proctoring for the exam. My job as a TA for the Math for Robotics course include preparation.

Mountain View, California, US

Developed new state-of-the-art model BAdMAN (Bagging Adversarial Multihop Attention Network) for Question Answer ranking task. The model was tested on 4 different datasets against several deep learning based baseline models. Applying for 2 papers and 3 patents based on the project.

I am working as a TA for two courses - Probabilistic Learning (undergraduate) at Computer Science Department, UCSD under Prof Lawrence Saul and Parallel Computing (graduate) at Physics Department, UCSD under Prof Michael Norman. I am giving two lectures on Introduction to Machine Learning for the Parallel Computing course. Other than that, my work as TA.

I am a Teaching Assistant for CSE 190 - Statistical Natural Language Processing under Prof. Ndapa Nakashole in Winter 2018 starting January. My duties include taking discussion sessions, preparing and verifying programming assignments for the course and helping out the professor with other course logistics.

Mclean, VA, US

The problem statement of Robocon 2015 is to make two badminton playing robots.

Mclean, VA, US

I worked in the Mobile NPS Digital Analytics Team.

My task is to supervise the all the projects being done currently at Models and Robotics Section, IIT Roorkee, guiding the teams whenever required - technically and otherwise.

Ahmedabad, Gujarat, IN

My project was to develop an automated welding table for welding of circular and cylindrical parts. Also, to visit all the transformer manufacturing plants to study various industrial considerations that go into manufacturing of various transformers.

Master Of Science - Ms, Computer Science

Bachelor Of Technology (Btech), Electrical And Electronics Engineering

Hsc, Elementary Education And Teaching