Game recommendation
What Game?
Python · Streamlit · scikit-learn
Find your next game with cosine similarity and IGDB data. A Streamlit app with seaborn visualizations to explore the recommendations.
GitHubApplied Scientist
Aamodit Acharya.
Currently at Ramp,
working on financial intelligence and reporting.
Statistics graduate
from the University of Waterloo.
New York & Toronto
Let’s connect
01 / About
It started with sports.
I’m Aamodit Acharya, a Statistics graduate from the University of Waterloo. I like messy problems, good people, and the point where a bunch of scattered ideas finally starts making sense.
Currently: Applied Scientist at Ramp, working full-time on financial intelligence and reporting.
My story started with arguing with friends about sports: hockey takes, F1 calls, and debates that needed more than vibes. I started pulling data, finding patterns, and using insights to prove them wrong. Somewhere in that chaos, I realized I liked turning messy questions into cleaner answers.
New York / Toronto
Aamodit Acharya
Interests
- search systems
- /
- experimentation
- /
- growth analytics
- /
- forecasting
02 / Experience
Experience
Data science, analytics, and the systems behind them.
2022 — Present
Working on financial intelligence and reporting at Ramp.
Current focus
Financial intelligence
Reporting
Earlier at Ramp · InternshipReimbursement automation
During my Ramp internship, I worked on reimbursement automation.
Data Science Intern
@ WealthsimpleBuilt production Marketing Data systems across Braze and Airflow: push-token observability, a daily attribute catalogue, and an AI-assisted DAG debugging agent.
Impact
25% higher marketing campaign engagement
75% less manual attribute investigation
30% less on-call debugging time
Tools
Data Science Intern
@ MashBuilt conversational search infrastructure. Improved answer relevance by 45% via RAG search pipelines. Expanded multimodal agent coverage by 50% and reduced generation hallucinations by 30%.
Impact
Better answer relevance (45%)
Multimodal search coverage (+50%)
Lower hallucination rate (-30%)
Tools
Data Science Intern
@ StatsigOptimized compute query pipelines for high-cardinality logs. Supported 5B+ daily events, cut Spark query times by 82% via HLL++ sketches, and pushed latency down by 25% on Iceberg tables.
Impact
Scalable experimentation metrics
Faster segmentation (25% latency cut)
Reliable high-cardinality modeling
Tools
Data Science Intern
@ TD BankDeveloped customer cancellation risk metrics on Hadoop databases using Impala SQL. Rewrote legacy processing scripts to reduce runtime by 97%, and engineered automated web scrapers on AWS EC2.
Impact
Customer retention signal (+22% retention)
Faster data refresh (97% speedup)
Underwriting market signal collection
Tools
Data Science Intern
@ TD BankContainerized forecasting scripts into Docker and Kubernetes microservices. Decreased forecasting process time by 67%. Built advisory utilization models using RidgeCV to boost closing rates by 25%.
Impact
Faster forecasting cycles (67% speedup)
Reliable containerized modeling pipelines
25% boost in closing-rate predictions
Tools
Data Science Intern
@ DesjardinsDeveloped monthly auditing SAS pipelines for fraud discrepancy reporting. Reviewed risk underwriting and pricing logic across 10 distinct variables to ensure alignment between Radar and R modeling frameworks.
Impact
Fraud discrepancy reporting (30% accuracy lift)
Segmentation risk review (10 alignment factors)
Radar to R script mathematical verification
Tools
Beyond work
Python drone navigation, YOLO detection, NumPy, PyTorch.
Financial planning, budgeting, events, and club support.
Robotics curriculum, mentoring, AutoCAD, C++, Python.
Credentials
Supervised Machine Learning: Regression and Classification — covering Linear Regression, Logistic Regression, Decision Trees, Random Forest, and Gradient Boosting.
Comprehensive assessment of probability concepts and their application in actuarial science, covering probability theory, random variables, and distributions.
Assesses understanding of fundamental financial mathematics — calculating values for cash flow streams used in valuing loans and bonds, asset/liability management, and investment income.
03 / Projects
Outside the day job.
Things I’ve built and questions I’ve explored.
Image captioning
What Image?
Python · PyTorch
An image captioning model pairing CNN image features with an LSTM to generate text. Built and trained in PyTorch.
GitHubMusic analysis
Drake vs Kendrick
R · kNN
A comparison of song popularity, play distributions, and artist features using exploratory data analysis and k-nearest neighbors classification.
GitHubVolatility modeling
GameStop Stock Analysis
R · Robust regression
Modeling GME volatility with linear regression, influence metrics, and robust regression via gradient descent to account for outliers and identify influential trading days.
GitHubSports statistics
Edmonton Oilers Analysis
R · MLE · MDE
Studying goal-scoring distributions with Poisson estimation, then comparing home and away goals using the Horvitz–Thompson estimator to assess bias, variance, and mean squared error.
GitHubPersonal automation
Sunday
Python · APIs · Automation
A notification pipeline that brings Google Calendar events, email updates, and real-time traffic together into timely, context-aware alerts.
GitHubSales forecasting
Mart Metrics
Python · XGBoost
A sales forecasting model using the Walmart Kaggle dataset, time-series decomposition, hyperparameter tuning, and cross-validation to predict sales trends.
In progress04 / Contact
Always happy to trade ideas, talk data, or hear a good hockey take.