Experience
-
Data Scientist
Adobe, San Jose, CA
Forecasting models and AI agents for marketing and finance teams, including agents that automate funnel insights, M&A signals and competitor tracking. Also an Adobe AI Ambassador, promoting AI adoption across teams.
Jun 2025 – Present1 yr 5 mos
-
Senior Data Analyst
BlackRock, San Francisco, CA
Led 20+ analytics and data science implementations for global clients, and built data cloud pipelines and generative AI tools for investment teams.
Aug 2021 – Jun 20253 yrs 11 mos
-
Data Analyst Intern
SAP, Houston, TX
Predictive models and ETL pipelines for customer success, and named SAP Best Intern 2021.
Oct 2020 – Apr 20217 mos
-
Data Science Intern
SkyGeni, San Jose, CA
Predictive models and ETL pipelines on go-to-market data for Fortune 500 clients.
Jul 2020 – Sep 20203 mos
-
Student Data Analyst
Northeastern University, San Jose, CA
Automated workflows in SQL, and was a Graduate Teaching Assistant for Predictive Analytics, mentoring 30+ graduate students.
Apr 2020 – Apr 20211 yr 1 mo
-
Data Analyst
ICICI Bank, India
SQL automation, a BigQuery data warehouse and Tableau dashboards, and named Top Analyst for A/B testing work on credit strategies.
Jun 2017 – Sep 20181 yr 4 mos
-
Data Analyst Intern
GIC Housing, India
Built PRiSE, a machine learning model for predicting loan defaults that reduced defaults by 28%.
Apr 2016 – Jun 20163 mos
Education
-
Stanford University, USA
Artificial Intelligence Program
-
Northeastern University, USA
MS in Data Science and Analytics
-
CIMP, India
MBA in Business Analytics and Finance
-
Patna Women's College, India
BS in Statistics and Finance
Certifications
- AWS Certified Cloud Practitioner
- Snowflake Data Engineering
- Databricks Certified Generative AI Engineer (pursuing)
Projects
Selected projects from my GitHub, grouped by where I built them. Open a row for the summary and a link to the code.
Stanford University
Artificial Intelligence Program, 2026-27
Independent implementations of ideas from the program, including search, constraint satisfaction, probabilistic inference, reinforcement learning, and deep learning for language.
Featured projects
Transformer language model Deep learning and NLP
A Transformer language model implemented in PyTorch, covering self-attention, multi-head attention, positional representations, masked language modeling, training and evaluation.
- Python
- PyTorch
- Self-attention
- Masked language modeling
Neural machine translation Deep learning and NLP
A neural machine translation system in PyTorch using sequence-to-sequence modeling, word embeddings, recurrent neural networks and attention mechanisms.
- Python
- PyTorch
- Seq2seq
- RNNs
- Attention
Adaptive Q-learning agent Reinforcement learning
A reinforcement-learning agent that learns optimal actions through Q-learning, balancing the tradeoff between exploration and exploitation.
- Q-learning
- Exploration and exploitation
Other projects
AI constraint scheduler AI algorithms
Generates conflict-free meeting schedules as a constraint satisfaction problem. It uses backtracking search with MRV and LCV ordering, forward checking and constraint propagation, and in the demo it schedules 4 of 4 meetings in 5 search nodes with no backtracking.
- Python
- Constraint satisfaction
- Backtracking
- MRV and LCV
- Forward checking
Intelligent route planner AI algorithms
Compares Uniform Cost Search and A* for optimal route planning on weighted graphs, using admissible heuristics, path reconstruction and search-efficiency benchmarking to show the trade-offs.
- Python
- Uniform Cost Search
- A* search
- Heuristics
- Graph algorithms
Probabilistic risk inference AI algorithms
A Bayesian network that estimates operational risk from observed evidence such as load, anomaly signals and service symptoms, computing posterior probabilities by enumeration.
- Python
- Bayesian networks
- Probabilistic inference
Adversarial game AI AI algorithms
A game-playing agent that uses minimax, alpha-beta pruning and heuristic evaluation for adversarial decision making.
- Minimax
- Alpha-beta pruning
- Heuristic evaluation
MDP resource optimizer Reinforcement learning
Optimizes dynamic compute capacity with Markov decision processes, using value iteration and policy iteration.
- Markov decision processes
- Value iteration
- Policy iteration
Neural dependency parser Deep learning and NLP
A neural dependency parser built with PyTorch, implementing transition-based parsing, feedforward neural networks and learned representations for syntactic dependency analysis.
- Python
- PyTorch
- Transition-based parsing
- Feedforward networks
NLP ticket classifier NLP and machine learning
A lightweight NLP text classifier that routes support tickets using TF-IDF features and linear classification.
- Python
- TF-IDF
- Linear classification
Northeastern University
MS in Data Science and Analytics, 2019-21
Featured projects
Lifetime value, churn and uplift Marketing analytics
Tracks engagement (DAU, MAU, ARPU) and revenue per user, calculates customer lifetime value from cohorts and retention curves, and builds churn and sales models. Uplift modeling and A/B tests measure what each campaign actually changed.
- Python
- Logistic regression
- Random Forest
- XGBoost
Revenue (ARR) and sales forecasting Forecasting
Forecasts annual recurring revenue to find growth patterns, seasonality and churn trends by segment. Also covers pricing elasticity, promotion impact, multi-touch attribution and CPA versus LTV.
- Python
- ARIMA
- Prophet
- LSTM
- CausalImpact
Amazon customer segmentation and recommender Marketing analytics
Builds customer-level features from transactions, clusters customers with K-means, profiles each segment, and recommends best-sellers to customers in the same cluster who haven't bought them yet.
- Python
- K-means
- Feature engineering
Other projects
Walmart sales price prediction Forecasting
Forecasts sales prices across stores and departments using time series features, and surfaces seasonal patterns, holiday effects and store-level demand differences.
- Linear Regression
- Random Forest
- XGBoost
- k-NN
Airbnb listings in New York City Exploratory analysis
Explores popular listings and the busiest hosts with Pandas, Seaborn and a geographic heatmap, and uses SQL (self joins, CTEs, procedures) to prepare data for a Tableau dashboard.
- Python
- SQL Server
- Tableau
Real estate price prediction Machine learning
Predicts house transaction prices with regression models compared on RMSE, R-squared and mean absolute error, using one-hot encoding for categorical features.
- XGBoost
- Lasso
- Random Forest
COVID-19 database analysis SQL
Cleans and explores 2020-2021 country-level case, death and vaccination records using joins, CTEs, temp tables, window functions and views.
- SQL Server
Nashville housing data cleaning SQL
Loads, cleans and restructures records of houses sold in Nashville between 2013 and 2019, using DML, DQL and DDL.
- PostgreSQL
Car price predictor Machine learning
Detects outliers and explores a dataset of more than 10,000 cars (fuel type, horsepower, transmission, mileage) before predicting price.
- Python
- Random Forest
Pokémon battle win prediction Machine learning
Predicts the likelihood of winning a battle and compares models. XGBoost on a reduced feature set had the lowest error, using only attack and speed.
- SVM
- Regression tree
- Random Forest
- XGBoost
- PCA
Boston truck permit analysis Exploratory analysis
Analyzes the city's daily export of moving truck permits, with modeling and visualization of the data.
- Python
- SciPy
- Pandas
- Matplotlib
Fortune 500 profitability and employee efficiency Exploratory analysis
Compares the largest US companies by region, industry and sector, and models the relationship between profit and headcount to find the KPIs behind efficiency.
- Python
- Random Forest
CIMP
MBA in Business Analytics and Finance, 2015-17
PRiSE: predictive risk scoring for loan defaults Finance and risk
A scoring engine that flags likely home loan defaulters from demographic, geographic, credit appraisal and loan-specific factors. It reduced loan defaults by 28% and the research won the Best Research Paper Award at GCBFR USA.
- Python
- SQL
- R
- Machine learning
All repositories are on GitHub.
Skills
- AI and LLMs
- HuggingFace
- RAG
- LangChain
- OpenAI
- Claude
- Generative AI
- OpenSearch
- Machine learning
- scikit-learn
- XGBoost
- PyTorch
- TensorFlow
- Keras
- OpenCV
- Prophet
- SARIMAX
- Programming and ETL
- Python
- SQL
- R
- dbt
- Airflow
- Alteryx
- Git
- Cloud and big data
- AWS
- Azure
- GCP
- Snowflake
- Databricks
- Spark
- Hadoop
- Hive
- BigQuery
- Oracle
- BI and visualization
- Tableau
- Power BI
- Looker
- AWS QuickSight
- Mode
- Sigma
- Excel
- Marketing and web analytics
- Adobe Analytics
- Google Analytics
- Salesforce
Achievements
Conferences and panels
I've attended conferences as a panelist and speaker on AI, technology and business topics, with a range of organizations.
-
2026
Artificial Intelligence & Sustainability Global Summit 2026 at AWS Builder's Loft, San Francisco, sharing perspective how AI is transforming customer experiences to reimagining enterprise workflows and decision-making.
-
2025
Society of Women Engineers 2025 Conference in New Orleans representing Adobe and discussing how data science and AI shape analytics and forecasting.
-
2022
Meta and BlackRock Youth Forum 2022 at San Jose to discuss the Data and Tech industry pathway.
-
2021
Analytics Round-table edition with Microsoft, BlackRock and Northeastern University with Azadeh Mobasher, Sr Data Scientist Microsoft and Amanda Welsh, Faculty Director at Northeastern University.
-
2020
Data Analytics in Business & Finance 2020 - Google, Costco and SAP panel discussion with Fidel Rodriguez, Head of Analytics and User Insights at Google and Trupti Shah, Data Analyst at Costco.
Hackathons and apprenticeships
I enjoy attending, and sometimes winning, hackathons and coding competitions. They're a good excuse to try new technology, meet people and drink a lot of free coffee.
- 2023
HackBLK, runner-up. Designed a commission-based auto-investing app using Keras to promote financial democratization.
- 2021
ServiceNow Accenture Hackathon, winner. Built an anomaly and threat detection model that reduced security vulnerability by 20%.
- 2021
Jeeva Health AI. Built a SciPy model and SQL to assess health datasets, improving preventative care strategies by 30%.
- 2020
Alicorn Venture. Developed Bitcoin price prediction SQL models that informed tech investment decisions.
- 2019-21
Google DSC. Prototyped an MVP and BI strategy for the AyyGo app, and built a go-to-market plan that reached $50K in revenue.
Publications
- 2017
Outstanding Research and Best in Session Award, 12th Global Conference in Business and Research 2017, USA. Impact of Statistical and Risk Variables on Non-Performing Loans. Read the paper
Outstanding Research Award, 2017
Best in Session Award, 2017
Presenting the research, 2017
- 2017
International Conference on Business Research 2017, India Trading in Electricity Derivatives: Evidence from India.
- 2016
4th Pan IIM World Management Conference 2016, IIM Ahmedabad, India Unifying Legal Framework in Debt Financing: Insolvency and Bankruptcy Code 2016.
Beyond work
Fun activities
- Badminton. My go-to way to switch off.
- Museums. I like wandering through them.
- Backyard gardening. Growing things at home.
- Skydiving. I recently did one of the world's highest tandem jumps, from 18,000 feet.
Reading
A few books I'd recommend, from history and memoir to AI and forecasting.
SapiensYuval Noah Harari
IkigaiHéctor García and Francesc Miralles
BecomingMichelle Obama
Human CompatibleStuart Russell
The Signal and the NoiseNate Silver