We just announced our $3M Pre-Seed. Watch our — launch video.
Cofounder & CEO at Kintsugi: saving businesses from sales tax in 7 clicks and 3 mins | MIT & Harvard MBA dropout | CS AI grad @ Stanford
With a focus and deep expertise in machine learning and all things AI, my journey has evolved into the role of Cofounder and CEO at Kintsugi, where I apply my technical prowess in Machine and Deep Learning to create vertical models that serve as indirect-tax engine that power Kintsugi AI.
My tenure as a Senior Machine Learning Engineer at Facebook was marked by significant contributions to ML filtering and Ranking models, enhancing the precision and effectiveness of Fb, Instagram and Messenger Stories product. The skills honed there now underpin my strategic direction at Kintsugi, guiding a team in crafting cutting-edge, AI-driven contextual models.
Chat with Clera and we'll introduce you to the right opportunities.
This profile is based on publicly available information. Pujun is not affiliated with or endorsed by Clera. Privacy policy.
Stanford University · Seasonal
Teaching graduate students Skills: Guest Lecturing · Public Lectures · Research Skills
Skills: Deep Learning · Machine Learning · Full-Stack Development
was working on Kintsugi as part of startup program Skills: CEOs · Full-Stack Development · Go-to-Market Strategy
Teaching assistant for CS221: Artificial Intelligence Skills: Deep Learning · Machine Learning
Ads growth: Dynamic Ads Creation and Targeting using Machine Learning Skills: Deep Learning · Machine Learning · Full-Stack Development
Working on Machine Learning and Deep Learning modeling efforts for better user targeting for Stories at Facebook. Skills: JavaScript · Deep Learning · Python · Machine Learning
Course assistant for CS276: Information Retrieval and Web Search. We look into questions like how to make a search engine, like Google. How to optimize it and come up with novel approaches to rank web pages. How to mine the internet for relevant information. We try to take a research approach, trying to look and push the boundaries of the state of the art techniques, while also staying grounded and coming up with algorithms/approaches that can scale really well in the industry and make a crucial difference in efficiency for these systems. Skills: Deep Learning · Machine Learning
At the Natural Language Processing (NLP) group. Skills: JavaScript · Deep Learning · Python · Machine Learning
Working on the Business Analytics and Modeling team as a full stack engineering intern. My summer project is to develop the data-pipeline monitoring system for finding potential loss of Kafka messages in Yelp's distributed producer-consumer model. Skills: Full-Stack Development · Machine Learning
In Spring 2015, I am also teaching a distance learning course along with regular 3 and 4 dimension calculus. As a part of that, all my lectures can be found on Mcgraw Hill Gatech portal and this semester, 80 more students are part of my online recitation. In total, I am responsible for 100+ students. Skills: Deep Learning · Machine Learning · Full-Stack Development · Data Science
In Spring 2014, I am also teaching a distance learning course along with regular calculus. As a part of that, all my lectures can be found on Mcgraw Hill Gatech portal and this semester, 30 more students are part of my online recitation.
Working with Greg Mayer, post-doc at School of Mathematics, Georgia Institute of Technology. Using statistical models to analyze online teaching interactions. Funded by Georgia Tech President’s Office Skills: Deep Learning · Machine Learning · Data Science · Full-Stack Development
Working on the core engineering team on dynamic ads and user heat maps Skills: Machine Learning · Full-Stack Development
In the spring semester, we took the idea of intelligent review to the whole different level by deciding to arrange all the questions differently for every student based on their strengths and weaknesses. For that, we chose to have a concept based approach and thus we used probability based models, cross correlation among different concepts and also developed a suitable mathematical trend ,based on the data for the last 6 years. With all this, we were able to classify all the concepts into 5 tiers (with Tier 0 being the most difficult) and all the corresponding questions accordingly (this was trickier since a challenging concept could have an easy concept associated to it.) After doing all this data mining and all the computation, we were able to get a list of all the concepts that were really important for any particular student and using this information, we decided to make binary 'questions' trees. For this, we had to make another model that took into account the ease and the effectiveness transition from one concept to the other. For this, we tried alot of series and compared our data with the past years data. After trying 10 or so models, we settled on one model and finally made the binary trees, thus totally redefining the whole system. This implementation will be used by Georgia Tech students starting Fall 2014.
As a lead teaching assistant, I am responsible for managing 200+ students who are taking advanced Calculus and Linear Algebra and coordinate with other TAs to make sure that everything runs smoothly. I conduct and weekly quizzes, homework, midterms and final examination. Also, I teach 35 of these students twice a week.
I got the wonderful opportunity to study at Oxford university, UK and further expand my international experience. For the first 6 weeks, we were accompanied by our professor on a journey through major countries of Europe and then for the last 6 weeks, we were students of the prestigious Worcester College at Oxford.
I was a full time student at Georgia Tech Lorraine, France and I visited more than 15 European companies while taking core ECE courses like ECE 2031, ISYE 3770 and Differential equations etc.
Enterprise2Empower · Part-time
I was the Vice President of Technology at Enterprise to Empower, an emerging student organization at Georgia Tech focused on entrepreneurship used LAMP architecture to maintain the organization marketing website and logged in experience Skills: Full-Stack Development
Undergraduate Teaching Assistant for MATH 1502 - Calculus II (Fall 2013) and MATH 1712 - Calculus and its applications (Fall 2012). Responsible for mentoring Undergraduates enrolled in the class, teaching and revising all important concepts,conducting and supervising quizzes, midterms and finals. In Spring 2014, I am also teaching a distance learning course along with regular calculus. All my lectures can be found on Mcgraw Hill Gatech portal
I was a tutor for Mathematics and Physics at Clough Undergraduate Learning Commons. A part of responsibility was to teach (both 1-to-1 and group) students, answer their doubts and kindle their interest in the subjects.
I worked with senior engineers of L&T and NDPL to simulate and analyze power grid system of North Delhi. We Used data mining techniques in MATLAB, C++, to determine potential flaws in power stations and by the end of the summer, we were successful in increasing the efficiency of the network by 40%
Doing research as a part of Vertically Integrated Projects (VIP) under Dr. James McClellan on Intelligent Tutoring System for ECE 2026 (ECE 2025) - Introduction to Digital Signal Processing for 3 semesters.
In Fall 2012, We designed and implemented algorithms, using N-gram analysis, for the Intelligent Tutoring System which currently in use at Georgia Institute of Technology. Our m*n list implementation for the concept base improved the student online question search system efficiency by 53%. In Spring 2012, we completely did a redesign and auto arrangement of concepts within the Intelligent Tutoring System using 3 level integration and I also proposed Instructor Assignment Creation system. Overall, our project was able to improve the system's efficiency by 29%. Skills: Full-Stack Development · Machine Learning
Bharat Heavy Electricals Limited
I worked with senior engineers to evaluate the feasibility of large scale usage of hydraulic energy for electricity. Our team reported changes to Bhakra Dam authorities to improve energy security by 21%
National Physical Laboratory (NPL)
I was fortunate and very lucky to be selected to be a part of CSIR youth leadership of Science program in India. As a result, I got the wonderful opportunity to study under and work with brilliant minds of India to analyse the various measuring scales and their manufacturing process. Later in the program, I was given the opportunity to study RF and FM waves and ultimately come up with an idea or approach to improve the wireless communication.
Started MBA but never finished. dropped out to cofound and pursue a startup. Currently based out of SF. Time at HBS was good but there is bigger fish to fry and company to build.
left MIT to work on my startup over the summer in the US which wasn't allowed by MIT Sloan as someone on F1 visa but with already a Masters degree. wanted to pursue an idea and staying at MIT wasn't worth not working on the idea so dropped out
Graduate school in Artificial Intelligence at the CS department. Graduated with a MSCS degree to work at Facebook fulltime as an ML engineer. kick-started ML team at FB stories.
Grade: 4.0 GPA I took core ECE classes while travelling to 15 European countries and having the time of my life as a 20yr old
Grade: 4.0 GPA Computer Engineering and Computer Science (Artificial Intelligence)
Grade: 4.0 major GPA Activities and societies: GT SAA, Eta Kappa Nu, Tau Beta Pi Computer Engineering and Computer Science (concentration in Artificial Intelligence) Skills: Full-Stack Development · Machine Learning · Mathematical Modeling
Selected as one of 20 high school students across India to conduct research at IIT Delhi at the age of 16. Earned this opportunity by ranking among the highest scorers in India in the CBSE Math, Science, and overall national examinations. Additionally, served as a research intern at the National Physical Laboratory as part of this program.
Activities and societies: Vice President at XINO Computer Science, Mathematics, Physics were favorite subjects. I could deal with Organic Chemistry. I hated inorganic Chemistry with a passion of thousand suns. was a nerd at school. what did you expect.
Claim it to keep it up to date, or request removal. We're happy to help either way.
Worked with 3 other TAs to conduct weekly 3-hour verbal, written and instructional lab sessions for Digital Design Lab at the School of Electrical and Computer Engineering, Georgia Institute of Technology. Helped 30 2nd and 3rd year ECE/CS students with conflict resolution, problem solving and help them develop an overall understanding on how digital electronics work.
In Fall 2013, we are developing a mobile application to be used by Tech students to access & finish homework on tablets. This will include a personalized flashcard app for students to review concepts & enhance student-instructor interaction. Skills: Deep Learning · Machine Learning · Full-Stack Development · PHP