Intuit

- Architect & developer of tool, for evaluating LLMs and Generative AI systems, used across Intuit. Contributor to Intuit’s Generative AI Platform covered by Forbes and Nasdaq
- Leading Generative AI experiences for TurboTax
- Led a team of Data Scientists, Machine Learning Engineers and Analysts to build and deploy a portfolio of NLP/Language models for customer feedback:
1. Transformer based text classification model which can recommend which customers need a follow up call alleviating responsibility from call centre agents. The model led to a 50% increase in call volumes, driving follow up calls for 650k customers leading to a 6% increase in customer conversion and 18% increase in yoy customer retention.
2. Transformer based sentiment model which can identify the true sentiment of customers from their feedback. Led to identification of at risk cohort of customers worth $6 Million completely unknown till now.
3. Unsupervised topic model which can identify top themes dynamically among millions of feedback verbatims and group them into those themes. Used to proactively identify software issues as well as influence product roadmap.
-  Architected and led development of a reusable, modular, scalable platform for training and deploying various NLP/Language models: from unsupervised to transformer based LLMs. The platform, with a novel config driven design, was reused by multiple other teams working on NLP/Language models. 
- Mentored 5 Data Scientists
- Setup vision, roadmap for the team after co-ordinating with senior leadership and led to successful execution of that vision.

Staff Data Scientist | Lead Machine Learning Engineer, NLP & Generative AI

- Enhanced churn model with new features and sampling strategies which led to a 4X increase in identifying churn among 50M customers. Model identified 5M at risk customers worth $500 Million for the company.
- Led the end to end development of a novel model to detect fear, uncertainty and doubt among active customers. Model drove various retention initiatives like offering chat with an expert, free upgrades, prioritization in inbound call queues, personalized marketing which led to $2 Million in incremental revenue

Senior Data Scientist | Machine Learning Engineer, Customer Growth & Retention

- Led the design and development of a credit risk machine learning model estimated to save $4 Million a year. 
- Designed, developed and shipped to production a machine learning model to gauge risk and determine payment options at checkout. The model runs on 10 Million customers and saves $2.4 Million a year. Co-owned the entire machine learning model life cycle.
- Worked with the office of the Senior Vice President to make every production machine learning model at Intuit explainable.

Senior Data Scientist | Machine Learning Engineer, Security, Risk & Fraud

- Designed, built and deployed to production a real time machine learning model to stop Account Takeovers across all Intuit products. It reduced Account Takeovers by 74% while flagging only 0.1% of the logins and improved on the previous solution by 6x. It scores >300 Million logins a month making it one of the largest models in the company. 
- Architected and built new features which improved multiple machine learning models in production in the security, risk and fraud domain. Account Takeover models  had precision lift of 2x and 2.3x while recall increased 1.8x and 1.5x. Account Sign up model had recall lift of 2.5x. Architected the codebase of the features to be modular, reusable, config driven and with 100% unit test coverage.
- Contributed a modular, reusable config generation module to Intuit AI's central Feature Store which generated more than 20,000 lines of config code saving Data Scientist time and significantly reducing errors due to manual entry.
- Introduced Explainable AI to all the production ML models of the team increasing transparency for stakeholders, uncover new behavioral patterns and help improve the models.

Data Scientist

University of San Francisco

Board Advisor for the AI Strategy program, School of Management at the University of San Francisco

Board Advisor, AI Strategy Program

Stanford University

Delivered guest lectures as part of the course - Building LLM Applications

Guest Lecturer

KDD conference 2019

The workshop received 32 paper submissions from 11 countries, the highest for an Intuit hosted workshop. At KDD, the workshop was attended by more than 240 people making it the most popular workshop at one of the world’s premier Data Science conferences. Stalwarts like Patrick Hall, Been Kim, Rich Caruana delivered keynotes and Intuit started off important conversations around model fairness, accountability and transparency at one of the biggest stages for AI.

Co-Chair of Explainable AI Workshop

Carnegie Mellon University

Worked on clearTerms which involved modelling the legal language used in terms and conditions policies of various websites and highlighting important statements to the users before they click 'I agree'. Training data was obtained by crowd sourcing. Worked with word embeddings and neural networks - both recurrent and convolutional to improve accuracy by 20%.

Tools used: AWS, python, numpy, tensorflow

Graduate Research Assistant

Google

Worked with the IoT team on a novel approach for indoor localization (indoor GPS) - using a distributed and wireless system of vibration sensors. My work was focused on data and signal processing from the sensors. Achieved a localization accuracy of 20 cm, beating the standard for indoor localization systems by 5 times.

Co-op for Graduate Project

Designed and built a Machine Learning model end to end (data mining, cleaning, feature engineering, train and test) for fraud detection in financial data. Also programmed the entire model into production. Statistics of built model: Accuracy – 90%, Recall – 95%

Tools and Technologies used: Spark, Hadoop, Hive, Scala, Python, NumPy, SciPy, Pandas, Scikit-learn, Matplotlib, SQL, Jupyter Notebook, AWS

Data Science and Machine Learning Intern

Worked on Giotto - IoT based Smart Campus Living Lab (Funded by Google) The project aimed at building the next generation IoT stack.

Worked on new paradigms in IoT - API tiers for sensor data access to ensure user privacy in IoT and location based access control.

Tools and Technologies used: Android, REST, Django, Python, JSON

Graduate Research Assistant 

IEEE SIGHT (Special Interest Group on Humanitarian Technology)

Responsible for Events and Strategy for this program aimed at encouraging and funding innovative technological solutions for under served communities. Program currently covering 35 countries, 1700 volunteers and impacting 18,000 lives. 

Formulated event structures and strategies leading to increase in effective engagements, increase in scale and number of events and reduction in financial cost of events. Developed and executed innovative volunteer engagement strategies for current as well as potential SIGHT volunteers.

Engaged with volunteer groups in 35 countries and organised events in Sri Lanka, Philippines, India, Indonesia, Japan, South Africa, Uganda, Germany, USA, Argentina, Mexico and Singapore.

Steering Committee Member

Deloitte India (Offices of the US)

Interfaced with the client on a $1.4 million project to understand business requirements and worked on the front end side of web applications using OpenUI5 Javascript library, HTML, CSS, Bootstrap. Also worked on data services to integrate with SAP backend.

Designed the architecture of an IoT solution improving the operational efficiency of an Oil&Gas client. 

Led an engagement involving 400 employees targeting under served communities.

Software Developer

Award winning Data Scientist, Machine Learning Engineer and Leader. Proven track record of leading AI teams and building multiple machine learning models end to end (design to production) which significantly improve business metrics - from reducing fraud, improving credit risk to improving efficiency of agents and increasing a product's conversion rate leading to millions in incremental revenue. I have architected modular, scalable ML systems and platforms reused by hundreds. Recent experience in building and deploying NLP and large language models. Interested in Explainable AI. Co-chaired a very popular workshop on it at KDD 2019.  I have worked with leading researchers on state of the art deep learning projects at Carnegie Mellon University (check out my projects section for more details). I also have experience working with big data tools (Hadoop, Spark, Hive) and building effective machine learning models end to end in production environments (check out my work at Intuit)
Strong Computer Science background. I worked on a Google funded project to build the next generation IoT stack. I have also developed mobile apps and IoT solutions which have been used by consumers from across the world, covered by media and presented at various conferences (check out my projects section for more details)
Strong leadership skills and passion for leveraging technology for good. I was one of the youngest Steering Committee members in IEEE history, helped manage a program engaging IEEE members in 36 countries to build solutions for local communities.

Taplio

Jineet Doshi's Linkedin Analytics

Want to drive more opportunities from LinkedIn?

Taplio

.css-1jgreyo{background-image:linear-gradient(284deg, rgba(39,43,72,1) 0%, rgba(33,47,183,1) 100%);color:transparent;-webkit-background-clip:text;background-clip:text;}Jineet Doshi's Linkedin Analytics

Want to drive more opportunities from LinkedIn?

Jineet Doshi's Linkedin Analytics