Production ML and GenAI,
Shipped
I design and deploy pragmatic data products that solve core business problems. Backed by a PhD and a bias for shipping, I partner across the revenue and operations stack, from B2B Sales and Marketing to RevOps and Sustainability, translating each team's competing priorities into a single product the whole business can stand behind.
Stakeholder Discovery & Needs Shaping
Aligning with the businessLatent Needs Research
Finding business white spaces
Problem Framing
Solving root causes, not symptoms
Outcome Clarity
Defining success metrics early
Pragmatic System Architecture
Designing for cost and scaleData Architecture
Structuring consumption-first schemas
Cost-Aware Design
Optimizing compute, latency, and cost
Built to Scale
Designing modular, extensible systems
Rapid Prototyping & Validation
Testing assumptions earlyPrecision Prototyping
Testing riskiest assumptions first
Stakeholder Validation
Getting early feedback from users
Ship Criteria
Agreeing on ship criteria before coding
Delivery & Scalable Execution
Shipping production-ready productsCross-Functional Leadership
Bridging business and engineering
Production-Grade Quality
Deploying robust, monitored systems
Measurable Impact
Proving business value with telemetry
Trusted Tools & Stack
Modern stack for data products & AI orchestrationSkills & Stack
Work Experience
Data Scientist - Applied AI & Customer Analytics
Choice Hotels International
- B2B Sales Lead Scoring Product: Architected an end-to-end AI product to score and rank prospective B2B customers. Reduced sales research time from an hour to minutes for teams managing 6,000 properties. Delivered a fully integrated Tableau application merging Salesforce, internal data, and third-party APIs. Engineered the underlying ML and LLM features to drive a unified sort score.
- Predictive Customer Lifetime Value (CLV) Engine: Developed a cohort-based predictive analytics product processing 50M+ transaction records. Delivered actionable intelligence that drove targeting strategies for $100M+ in seasonal marketing campaigns.
- Vision-AI Brand Compliance Product: Architected an automated GenAI application to evaluate 300,000 marketing images across 6,000 franchised properties. Replaced a manual audit, immediately cutting $45K in operational costs. The deployed product is projected to drive a $5M revenue lift over the next year.
- Enterprise Entity Resolution Service: Shipped a multi-stage data matching product to reconcile external vendor data against 60,000 US hotel inventory. Upgraded legacy systems by integrating geospatial logic and LLM-based vector similarity for edge cases. Boosted successful match rates from 50% to 96%. Deployed the scalable architecture into production via AWS Glue.
- GenAI Climate Risk Application: Built a first-of-its-kind intelligence application for area directors managing 6,000 properties for climate risks and sustainability opportunities. Translated complex open-source climate data into intuitive KPIs. Integrated generative AI (Claude via Bedrock) to deliver actionable, property-specific summaries for non-technical users.
Lead Transportation Networks Data Scientist
ICF International
- Statewide Traffic Reliability Product: Built a comprehensive Python analytics engine to quantify travel time reliability for the Virginia Office of Intermodal Planning and Investment. Processed and scored millions of individual road segments across the state network. Partnered directly with leadership to define product requirements, translating complex traffic methodology into a deployed data asset used for statewide planning.
- Urban Mobility Impact Simulator: Architected a Python and QGIS spatial modeling product for NYSDOT to evaluate infrastructure changes along the Bronx Highway corridor. Designed an isochrone-based scenario engine to simulate and quantify how proposed access restrictions would affect resident transit accessibility. Translated complex spatial models into executive-ready metrics, directly guiding major urban planning decisions.
- National GTFS Evaluation Service: Engineered an automated data quality application using Python to audit transit network geometry across nearly 3,000 regional operators. Utilized spatial buffering techniques in GeoPandas to identify routes detached from underlying road networks. Provided FHWA stakeholders with a prioritized triage system, transforming fragmented public data into an actionable resource allocation strategy.
Transportation Systems Data Modeler
C&M Associates
- Predictive Forecasting Pipeline: Built the data ingestion pipeline powering Traffic and Revenue models for the I-495 extension. Used Python to automate data cleaning, missing value imputation, and outlier detection across over 500 traffic data collection stations to calibrate long-term simulation models.
- Revenue Scenario Simulator: Developed forecasting tools to drive the 30-year financial and strategic plan for the Dulles Toll Road. Modeled complex demand sensitivities and toll rate variations using Python and Excel, delivering actionable models directly to stakeholders.
Data Science Graduate Research Assistant
George Mason University
- Predictive Modeling & Data Architecture: Engineered automated pipelines to collect traffic data via API integrations and web scraping, building a centralized spatial inventory. Cleaned and structured disparate datasets to run machine learning models, mathematical algorithms, and Monte Carlo simulations. Tested complex infrastructure scenarios to drive the quantitative methodology for published academic research.
- Executive Intersection Analytics Dashboard: Built a public-facing Tableau application visualizing complex traffic flows across over 200 Fairfax stations. Parsed unstructured, irregular sensor text logs into a clean relational database. Engineered custom KPIs for seasonal variation and data reliability, delivering an end-to-end analytics product for executive decision makers.
Education
Personal Projects
Interactive systems, models, and analytics dashboards built to explore public data and showcase data science solutions.
Always open for collaborations!
Always ready for comments, suggestions, and collaborations on a product. Email me to pitch your idea!
Games Showcase
Interactive voxel experiences and game experiments.
Atabak & Alma Mardan
Games on this page were developed in collaboration with Product Owner Alma Mardan.
A fun voxel game co-developed with my 7-year-old daughter Alma (Product Owner), and fully vibe coded.
An addictive 8-bit retro runner game by Alma. Select your champion and jump over obstacles to keep the brain jumping!
Atabak Mardan, PhD
PhD in Transportation Systems & Operations Research
I received my PhD in Transportation Systems from George Mason University, with my doctoral thesis focusing on the modeling and optimization of dynamic pricing toll roads (Express Toll Lanes). My research covers traveler route choices, traffic simulation under dynamic fees, and travel time reliability. I specialize in designing and deploying operations research frameworks, statistical learning, and spatial analysis tools.
I am highly passionate about transport economics, traffic flow physics, and network optimization, and I am always ready to collaborate, discuss new ideas, and consult on related data products.
Google Scholar Profile
Selected Publications & Research Output
| Publication Title | Citations | Year |
|---|---|---|
Insights & Activity
Thoughts on Applied AI, Data Engineering, and Building Products.
Get in Touch
Let's discuss predictive systems, operations research, spatial modeling, or data architecture.
Let's build the next generation of data products.
Have an idea for a project, a suggestion, or a potential collaboration? Send a message below or use the direct email link as a backup.
Message Sent!
Thank you for reaching out, . I have received your message and will get back to you within 24 hours.