Ram Kaashyap Prasanna
Quantitative Researcher · Consumer Credit & Structured Finance· New York, NY
I'm a quantitative researcher at the intersection of credit modeling and structured capital deployment. At Edge Focus, a credit fund investing in consumer unsecured loans, I build the probability-of-default and prepayment models behind its IRR calculations and investment decisions, and design and run the lending programs that put capital to work across originators, warehouse lenders, and capital investors. Before moving to the buy side, I spent nearly three years at Open Lending building in-house credit risk infrastructure, after earlier roles at Cottonwood Financial and AryaNow.
At a glance · hover for details, click to jump to an entry
- Work
- Education
- Personal projects
Skills · tap one to see where it shows up
- Credit & structured finance
- Statistics & ML
- Programming & data
- Data sources
Now
Apr 2026 – Present
Build probability-of-default and prepayment models whose output feeds directly into IRR calculations and drives fund-level investment decisions.
- Design and structure capital deployment programs across the full counterparty stack (loan originators, warehouse lenders, capital investors), and manage active lending programs end-to-end, from inception through performance monitoring and counterparty relationships.
- Audit the buyer-protective triggers in the loan sale agreement (collateral performance tests, Event-of-Default escalation, sponsor covenants, concentration limits), and build a Python trigger-monitoring dashboard driven by a credit-model roll-forward to forecast trigger headroom.
- Analyzed an originator's re-age program and quantified its downstream effects on cumulative gross loss curve shape, IRR, WAL extension, and under-amortization risk.
- Lead credit-risk feature engineering on TransUnion tradeline data, including telecom, rental, and utility segments, through a zero-leakage signal discovery engine; benchmarked the proprietary score against the bureau score with Kendall's tau across months-on-book vintages.
- Translate model output into rate-setting and pricing recommendations for originators; review credit-guideline changes for volume and loss impact; research alternative-data vendors (Equifax, TransUnion, Experian, LexisNexis) for second-look and turn-down programs.
- Authored an installable Python credit-modeling toolkit (WoE/IV, OptBinning with monotonic constraints, a LightGBM training wrapper, PSI, KS/Gini/AUC, PDO score alignment) used as shared infrastructure by the team.
- Review team research, forecasts, investment memoranda, and code; collaborate with external partners on deal evaluation and due diligence.
2026
Oct 2026
Five hypotheses about Flip 7 strategy, a full simulation of the game, and 200,000 games. There is no magic number of cards. The best rule reads the odds of the next card, and it is the same rule at every table size.
Oct 2026
Hypothesis: on the grid, crossing the avenue when the light allows and turning uptown when it doesn't beats the plain L. A simulator with NYC signal timing says yes, by about 50 seconds on a 13-minute walk, if you are patient at the corner.
Oct 2026
My Mac mini stays on at home and I use its full desktop from my iPad, on Wi-Fi or cellular. Tailscale, Jump Desktop, a fallback or two, and a watchdog.
Oct 2026
My Flip 7 scorekeeper now has an Apple Watch app. I tap the cards on my wrist, the phone (or the TV) shows the race to 200, and every finished game is saved round by round.
Sep 2026
A nine-turn game for 2–6 people where you fight over a hex map with dice you build yourself. It started as a phone app for game night; now it runs anywhere with a browser. Early playtest build, rules still moving.
Aug 2026
An e-ink screen by the door that shows the weather and, for the two stations near me, the next trains with the time I have to leave to catch them. Code and setup guide included.
Aug 2026
Rebuilding the word2vec model and the class, gender, and morality dimensions from Kozlowski, Taddy and Evans (2019), checking them against survey data, and putting words in an interactive 3D space.
Oct 2023 – Apr 2026
- Senior Data ScientistDec 2025 – Apr 2026
- Data ScientistOct 2023 – Dec 2025
Built the company's first in-house probability-of-default scorecard for near-prime to prime borrowers: 15% higher rank-order lift (AUC) than the generic bureau score and 7% higher than the prior internal model.
- Ran the full build on TransUnion and LexisNexis data (cleaning, feature engineering and selection, tuning, robustness testing), benchmarking logistic regression against XGBoost, LightGBM, and CatBoost.
- Took the model through production and governance with IT, Compliance and Legal, TransUnion, and a third-party Fair Lending auditor for FCRA best practices.
- Led development of the second-generation risk scoring model for the core portfolio.
- Found double the early delinquency on applications carrying "Credit Builder" tradelines, driving a credit-tightening policy change and a custom TransUnion variable.
- Implemented the first CI/CD process in Databricks (PySpark) to monitor production scorecard health with PSI and automated alerts, and set up champion/challenger and A/B testing procedures.
2025
2023
May 2023 – Oct 2023
Developed in-house probability-of-default and fraud models (logistic regression, XGBoost) optimized for cumulative cash balance / NPV and minimal realized losses.
- Built an end-to-end vintage performance monitoring pipeline: SQL Server and MariaDB sources consolidated, Python ETL on Databricks, and Power BI dashboards for KPIs and early-warning indicators refreshed up to twice a day.
- Built an Azure Data Factory pipeline with PySpark ETL, creating tables used by the Credit Risk and Marketing Analytics teams.
- Added Power BI population stability dashboards with population-change alerts.
- Ran A/B tests and cohort analyses in Python and SAS for senior management, and worked with IT, Product, Marketing, and Compliance to implement credit risk policies and underwriting criteria for new products.
May 2023
A working note on building an autonomous AI agent with GPT-4, LangChain, and AutoGPT.
May 2020 – May 2023
- Data ScientistJan 2021 – May 2023
- Product OwnerJan 2021 – May 2023
- Technology AnalystMay 2020 – Dec 2020
Built a discrete-event simulation (SimPy) of on-demand clinical operations, with stochastic patient arrivals, staff schedules, and fatigue, producing queue, throughput, revenue, staffing-cost, and profit distributions to guide staffing and capacity decisions.
- Delivered competitor research and the financial modeling behind the direct-to-consumer pricing plan, presenting business cases and ROI to the executive team.
- As Product Owner, led 2 business analysts, 2 UX experts, 2 UI developers, and 15 software developers through agile sprints on the Urgent Psychiatry product, with weekly executive presentations.
- Built a GPT-3 FAQ chatbot and proofs of concept including a video questionnaire with facial-emotion and sentiment analysis, a smart EHR system, and live driver's-license OCR.
- Served as Data Protection Officer, responsible for data integrity and security.
- Traveled to Kanyakumari, India to run an orientation for the AryaNow India team.
- As Technology Analyst, shaped the original telehealth psychiatry product flow with the psychiatrists on the executive team, UX patient interviews, and the COO and CTO.
Aug 2022 – May 2023
Year-long applied machine learning program. GPA 3.9.
- Data wrangling, exploratory analysis, visualization, and preprocessing
- Regression and classification models, ensemble techniques, and feature selection
- Neural networks, computer vision, and natural language processing
- Statistical learning, recommendation systems, and model deployment
Apr 2023
Notes on the required reading for my credit risk team: the seven stages of building a credit risk scorecard, plus the questions the book raised for me.
Apr 2023
Notes from Arize's online event on LLMs and ML observability, with talks from OpenAI, HuggingFace, Spotify, Etsy, LangChain, BabyAGI, and more.
Jan 2023
Introducing a project to predict Tesla's stock price with recurrent neural networks, combining traditional market data with cultural signals scraped from Reddit and Twitter.
2022
Dec 2022
A step-by-step walkthrough of building and validating a linear regression model on the 1990 California census housing dataset, ultimately explaining 64.4% of the variance with three degrees of interaction terms.
2020
Mar 2018 – Jul 2020
Taught K-12 students everything from addition and subtraction to Calculus II, and helped run events for kids in the program.
2016 – 2020
- President's List· 4.0 GPAAug 2019
- Dean's List· 3.75+ GPAAug 2018
- Senator, College of Science
- President, Model United Nations
2018
May 2018 – Aug 2018
Supported the Customer Management & Product Portfolio Management service delivery team, delivering projects in C#/.NET, SQL, and JavaScript/Angular.
Earlier
- Inter-State Robotics Competition· Winner, junior year
- 1st Place in Speech· Academic Decathlon of Texas, prepared speech
- 3rd Place in Interview· Academic Decathlon of Texas, three-judge panel
Certifications
- Machine Learning Model DevelopmentUplimit · Jun 2025
- Machine Learning at ScaleUplimit · Jun 2025
- Market Basket Analysis in PythonDataCamp · Apr 2023
- Machine Learning for Marketing in PythonDataCamp · Apr 2023
- Credit Risk Modeling in PythonDataCamp · Mar 2023
- HIPAA & Data SecurityAccountable HQ · Sep 2022
- Google Finance Data Analyst Professional CertificateGoogle · Aug 2022
- Google Data Analytics SpecializationGoogle · Aug 2022
- Healthcare Data Security, Privacy, and ComplianceJohns Hopkins University · Aug 2022