Summary
Overview
Work History
Education
Skills
Websites
Accomplishments
Workingpapers
Projectsexposure
Training
Timeline
Generic

Yagha (Josh) Joshi

McLean,VA

Summary

A capable Quantitative Analyst possessing a strong foundation in statistics and quantitative economics, empowered to tackle complex analytical challenges. An adept Data Scientist skilled in constructing and implementing predictive models by leveraging machine learning and deep learning methodologies. The exemplary educator, with the ability to embody inspiration, showcases exceptional teaching abilities and an undeniable history of success.

Overview

15
15
years of professional experience

Work History

Senior Quantitative Analyst

Federal Home Loan Mortgage Corporation (Freddie Mac)
01.2023 - Current
  • Models Managed: 1
  • Housing Price Appreciation (HPA) 2
  • Multifamily Analytical Credit Solutions (MACS) 3
  • Multifamily Current Expected Credit Losses (CECL) 4
  • Multifamily Collateral Financial Data Set (CFDS) 5
  • Operational Loss Aggregation Model 6
  • MF Market Tiering Model
  • Multifamily Analytical Credit Solutions (MACS) - Detected that the Loss Given Default (LGD) positive loss likelihood model has poor back-testing performance on
  • Freddie Mac defaulted loans. - Identified and suggested to the Multifamily Modeling team to treat an outlier in the LGD expense model data, improving the model's performance by 9%. - Identified that the LGD principal loss percentage component model under-forecasts principal loss by around 10%.
  • After adjustment, the loss was reduced to 1.7%. - The LGD principal model under-forecasts principal loss for defaulted loans with a one-quarter after default where
  • LTV ≥ 140%, presenting a notable risk since these are high loan to value (LTV) loans.
  • Identified a misspecification issue in the Housing Price Appreciation (HPA) Model, potentially impacting loss estimation by $200-800 million due to varying Expected Cost Default (EDC) impacts from different constraint implementations.
  • As a member of the Enterprise Model risk (EMR) team, I urged model owners to rectify the issue to mitigate these
  • EDC impacts.
  • Successfully identified and resolved an issue concerning automated input data used for final valuation in Collateral
  • Financial data Set (CFDS), which posed a risk of inaccurate financial valuations
  • Given the $7 billion portfolio, even a minor 1% price movement could potentially impact the income statement by approximately $70 million.
  • Developed a mechanism to mitigate a $49 million loss in Expected Default Costs (EDC) for the MF Market Tiering
  • Model, which was caused by prediction errors in vacancy and rent growth at the MSA level.
  • Advised the Model Owner to conduct a root cause analysis for the increase in the 9-quarter average forecast error for frequency prediction in the Operational Loss Aggregation Model, which has risen to 29% from 21% last quarter
  • The expected loss is estimated at $105M, based on data from 2010Q2 onwards, adjusted for a 4% inflation rate in 2023, contributing to increased historical losses in dollar terms.

Quantitative Analyst, AVP

State Street Bank & Trust
06.2021 - 12.2022
  • Re calibrated Muni Model that model demonstrated a significant 6% enhancement in overall predictive performance.
  • Played an integral role in implementing triggers that alert stakeholders to significant changes in customer behavior or engagement patterns, helping to identify and retain 3% of potential clients who may have otherwise left.
  • Played a pivotal role in implementing automated anomaly detection algorithms in an AML Customer Risk Rating Model, aiding financial institutions in swiftly identifying suspicious patterns and transactions linked to money laundering.
  • I played vital role to develop a Commercial Real Estate (CRE) Climate Risk Model 1
  • Assisted to enhance climate scenario analysis and stress testing methodology 2
  • Prepared climate risk assessment model scorecard
  • Stable Value Wrap (SVW) Model (Annual run for CCAR and quarterly for Basel)
  • Conducted Modeling, Forecasting, and Analytics for CCAR, DFAST, CECL, IFRS 9, and Basel Using Empirical Time
  • Series Econometric Methods

Instructor of Record

Southern Illinois University Carbondale
08.2017 - 05.2021
  • Money and Banking, History and philosophy of world's economic system
  • Econometrics I, Mathematical Economics Intermediate Microeconomics, Macroeconomics

Teaching Assistant

University of North Texas
08.2015 - 07.2017
  • Mathematical Economics, Basic Statistics

High School Mathematics Teacher

Shuvatara School
08.2009 - 06.2015
  • Mathematics

Education

MS - Computer Science

Southern Illinois University Carbondale

PhD - Quantitative Economics and Econometrics

Southern Illinois University Carbondale
05.2021

MS - Economics

University of North Texas
07.2017

MA - Economics

Tribhuvan University
07.2008

BS - Mathematics

Tribhuvan University
07.2003

Skills

  • Python
  • R
  • Stata
  • SAS
  • Matlab
  • SQL

Websites

Accomplishments

  • Thomas and Chaney Chung Endowed Scholarship, Southern Illinois University Carbondale, Fall 2020
  • The Augusta Jimmy Auerbach Endowed Memorial Scholarship, Southern Illinois University Carbondale, Spring 2019
  • Thomas and Chaney Chung Endowed Scholarship, Southern Illinois University Carbondale, Spring 2018
  • Texas Public Educational Grant, University of North Texas, Spring 2017

Workingpapers

  • Telecommunications and Innovation: A Global Study of Broadband, Telephone, and Mobile Cellular Access
  • Broadband Connectivity and Scholarly Communication: Enhancing Research Output and Impact
  • Impact of Broadband on Employment: Micro-economic Perspective.
  • Macroeconomic Policies in a Multiparty Democracy: A Game-Theoretic Approach

Projectsexposure

  • Applied Regression Analysis: Factors Influencing Run-Scoring in the Indian Premier League: Insights and Evidence
  • Security in Cyber Physical System: Simulation of Modbus/TCP- Base Industrial Control System
  • Applied Multivariate Statistics: Predicting Breast Cancer Types Beyond the Molecular Level: A Multi-Modal Machine Learning Approach
  • Deep Learning: Predicting Stock Prices with Long Short-Term Memory (LSTM) Networks and Time Series Models
  • Data Mining and Big Data Analysis: Cluster analysis and display genome-wide expression patterns using My own data.
  • Advanced Python Programming: Prospects for Economic Expansion: Predicting Nepal's GDP Growth

Training

  • Time Series Analysis in Python and R
  • Quantitative Analyst in Python and R
  • Data Scientist in Python and R

Timeline

Senior Quantitative Analyst

Federal Home Loan Mortgage Corporation (Freddie Mac)
01.2023 - Current

Quantitative Analyst, AVP

State Street Bank & Trust
06.2021 - 12.2022

Instructor of Record

Southern Illinois University Carbondale
08.2017 - 05.2021

Teaching Assistant

University of North Texas
08.2015 - 07.2017

High School Mathematics Teacher

Shuvatara School
08.2009 - 06.2015

MS - Computer Science

Southern Illinois University Carbondale

PhD - Quantitative Economics and Econometrics

Southern Illinois University Carbondale

MS - Economics

University of North Texas

MA - Economics

Tribhuvan University

BS - Mathematics

Tribhuvan University
Yagha (Josh) Joshi