Akshit Arora

Product Manager

Daly City, California, United States10 yrs 3 mos experience

Most Likely To SwitchAI ML Practitioner

Key Highlights

Expert in scalable Generative AI systems
Passionate about language and speech technologies
Published research in cognitive science and AI

Stackforce AI infers this person is a SaaS expert specializing in AI and machine learning solutions for enterprise applications.

Contact

akshit.arora1995@gmail.com LinkedIn

Skills

Core Skills

Artificial Intelligence (ai)Deep LearningTeachingMachine LearningSoftware EngineeringAugmented Reality (ar)Web Development

Other Skills

Amazon Web Services (AWS)Apache BeamC#Cognitive ScienceConvolutional Neural Network (CNN)Convolutional Neural Networks (CNN)Data ScienceData StructuresDockerExperimental DesignGithubGoogle Cloud PlatformKubernetesLinuxMentoring

About

As a Senior Solutions Architect at NVIDIA, I help large enterprises train and deploy Scalable Generative AI systems at scale, using NVIDIA's GPUs and software. I have a passion for language, speech and audio data, and I specialize in large language models (LLM) and text-to-speech (TTS) algorithms and applications. I have contributed to multiple TTS projects, such as scaling NVIDIA’s Multi-speaker Multi-lingual TTS systems with Zero-shot TTS to Indic Languages. I have a Master of Science in Computer Science from the University of Colorado Boulder, where I explored and published papers on various topics, such as cognitive science, climate prediction and AI for education. My academic background and professional experience have taught me how to apply deep learning to solve real-world problems and deliver impactful solutions. I am always eager to learn new skills and technologies, and to collaborate with other experts and innovators in the field of conversational AI. https://aroraakshit.github.io/

Experience

10 yrs 3 mos

Total Experience

2 yrs 2 mos

Average Tenure

6 yrs 9 mos

Current Experience

Nvidia

2 roles

Senior Data Scientist - Cloud Services

Promoted

Jun 2022 – Present · 4 yrs · San Francisco Bay Area · Hybrid

Building Scalable Generative AI Systems 🪄

Artificial Intelligence (AI)Text-to-SpeechDeep LearningMachine LearningPythonNVIDIA GPUs

Data Scientist - Consumer Internet

Sep 2019 – Jun 2022 · 2 yrs 9 mos · San Francisco Bay Area · Hybrid

Saga education

2 roles

Software Engineer Intern

Jan 2019 – Feb 2019 · 1 mo · Boulder, Colorado, U.S.A. · On-site

Likely my one and only official software engineering experience :)
Implemented a server-less pixel tracking based business intelligence pipeline using Google Cloud Platform (Pub/Sub and StackDriver Logging) and Apache Beam.
Built Ruby based SDK and CLIs to manage lifecycle of analytics pipeline on GCP’s BigQuery and Data Studio.

Google Cloud PlatformRubyApache BeamSoftware Engineering

Machine Learning Intern

May 2018 – Aug 2018 · 3 mos · Boulder, Colorado, U.S.A. · On-site

This experience made me realize how hard it was to deploy ML/DL in production -- which kicked off my endeavor to NVIDIA as a solutions architect.
Created an ensemble deep learning model for detecting student activity on their digital scratchpad (doodles vs mathematics) in a live classroom using recurrent neural network (RNN) and convolutional neural network (CNN).
Used Google's Doodles dataset and achieved 70% accuracy using TensorFlow (TF) Estimator API.
Further fine-tuned the model to include mathematical categories (coordinate planes / fractions etc.) by creating synthetic data from scratch using tools such as Pandas, Jupyter, MatplotLib, Scikit-Learn and Node.js.
Deployed the deep learning model in production including TF Serving, Docker & TF JavaScript.

TensorFlowRecurrent Neural Networks (RNN)Convolutional Neural Networks (CNN)Machine Learning

University of colorado boulder

3 roles

Research Assistant - Climate and Machine Learning Boulder (CLIMB)

Aug 2018 – Dec 2018 · 4 mos · Boulder, Colorado, U.S.A. · On-site

And after a bad experience with TensorFlow, I fell in love with PyTorch! (till date)
Created predictive models based on Convolutional Neural Network (CNN) for predicting climatic variables like precipitation and temperature using PyTorch.
Used time series data from predictors such as air temperature, sea surface temperature, wind velocity, sea level pressure and humidity.
Worked with Prof. Claire Monteleoni and Dr. Moumita Saha

PyTorchConvolutional Neural Network (CNN)Machine Learning

Research Assistant - MozerLab

Jan 2018 – May 2018 · 4 mos · Boulder, Colorado, U.S.A. · On-site

Yes, I worked on TensorFlow in 2018, and discovered it was so so broken :)
Trained and evaluated RNN based deep learning models for deep knowledge tracing using TensorFlow, Scikit-Learn, MatplotLib, Pandas and Keras. Compared their performance with Bayesian knowledge tracing approaches.
Presented a research paper on tracing student’s knowledge state on online tutoring platforms using Recurrent Neural Networks (RNN), at Educational Data Mining (EDM) 2018 conference.
Worked with Prof. Michael C. Mozer and Dr. Shirly Montero.

TensorFlowScikit-LearnRecurrent Neural Networks (RNN)Machine Learning

Graduate Teaching Assistant - College of Engineering and Applied Science

Oct 2017 – May 2019 · 1 yr 7 mos · Boulder, Colorado, U.S.A. · On-site

1. Introduced students to Computer Science via Python
2. Introduced students to data structures via C++ and Python
3. Taught the science of social statistics via R
held weekly labs teaching students about Python, Data Structures and Statistics (using R programming language)
created exams and assignments for the course
held office hours to help students with their doubts
held grading interviews and evaluating assignment/exam submissions

PythonData StructuresStatisticsTeaching

Indian institute of technology, mandi

2 roles

Research Assistant - Applied Cognitive Science Lab

Jan 2017 – Jun 2017 · 5 mos · Kamand, Himachal Pradesh, India · On-site

Designed and developed a search-and-rescue game using marker-based Augmented Reality (AR) that uses gestures to capture virtual objects in real world. Used C#, Vuforia, Unity 3D, RESTful API and packaged the experiment as an Android application.
Designed first-person-shooter scenario using marker-less GPS-based AR (similar to Pokemon Go) and evaluated libraries including Kudan AR, AR Toolkit, Microsoft HoloLens, Google Tango.
Performed experiments for cognitive profiling.
Worked with Dr. Varun Dutt and Akash Rao in collaboration with Institute of Nuclear Medicine & Allied Sciences (INMAS), Defence Research and Development Organisation (DRDO), New Delhi.

C#VuforiaUnity 3DAugmented Reality (AR)

Research Assistant - Applied Cognitive Science Lab

Jun 2015 – Dec 2016 · 1 yr 6 mos · Kamand, Himachal Pradesh, India · On-site

Built a serious game for creating landslide prevention and mitigation awareness people living in landslide-prone areas.
Designed a web-based interactive landslide simulator using PHP, Bootstrap and MySQL and studied the effect of affective feedback in the simulation. Performed Monte-Carlo simulations using Microsoft Excel.
Presented at a Springer conference called Applied Human Factors and Ergonomics (AHFE) in 2016 and published in Natural Hazards and Earth System Sciences journal.
Worked with Prof. Varun Dutt and Pratik Chaturvedi in collaboration with Defence Terrain Research Laboratory (DTRL), DRDO, New Delhi

PHPMySQLMonte Carlo SimulationWeb Development