cv
Background Information
Full Name | Novin Ghaffari |
Languages (natural) | English, Persian, Spanish (WK), Arabic (WK) |
Languages (computer) | Excel, LaTeX, Python, R, SQL |
Education
-
2019
PhD - Statistics
University of Texas at Austin
- Conducted cross-disciplinary mathematical research on ‘optimal transportation’ topic
- Developed several results and algorithmic applications for Wasserstein barycenters
- Applied optimal transport concepts to consensus modeling project with Sandia National Labs
- Advisor - Dr. Stephen G. Walker
-
2014
Masters - Statistics
University of Texas at Austin
- Researched stable distribution theory and implemented MCMC and FFT routines for inference
- Worked on product defect classification project with Freescale, Inc.
- Advisor - Dr. Carlos Carvalho
-
2012
BA - Plan II Honors
University of Texas at Austin
- Performed well in cross-disciplinary honors program with student profile akin to ivy league
- Wrote undergraduate thesis on advanced topics in mystical philosophies
- Advisor - Dr. Robert King
-
2012
BBA - Finance, Minor - Supply Chain
University of Texas at Austin
- Studied corporate finance, supply chain management, and general business
- Participated in Brass Rings practicum course with Freescale, Inc.
-
2012
BS - Applied Mathematics
University of Texas at Austin
- Completed math degree in 1 year in preparation for graduate school
Industry Experience
-
Jan 2021 - present
Head of Data and Analytics
Nulixir, Inc.
- Conducted background research and analysis covering business intelligence, product formulation, and marketing, and clinical scientific data for nutraceuticals
-
Business intelligence activities involved
- Researching macroeconomic conditions, particularly conditions affecting the nutraceutical industry
- Investigating competitor business strategies, technologies, products, and product claims
- Writing technical reports summarizing competitor technology
-
Product formulation research involved
- Researching evidence for the clinical effects of key active ingredients
- Reporting potential adverse effects, allergenicity, and toxicity
- Finding potential product formulation difficulties (e.g., instability, poor taste, etc.)
-
Marketing-related activities involved
- Analyzing Google search and social media data to determine nutraceutical opinion trends
- Scraping industry data to understand trends in nutraceutical sales
- Converting scientific data and evidence in simple marketing statements
- Subject matter expert in statistics and data science assisting several projects and business functions
-
Developed a rudimentary database for research and formulation purposes
- Database was developed and implemented in MySQL
- Data schema design was implemented using sqlalchemy in Python
-
Developed API access tools in Python to assistant lab scientists with research
- Developed wrapper functions for easy access to PubChem’s PUG REST API via pubchempy
- Designed functions for accessing JSON data from Pubchem’s PUG_VIEW API
- Implemented functions for downloading and viewing spectroscopy output images for calibrating lab equipment
-
Supported internal clinical trials with several data functions
- Provided experimental design methodology expertise for organizing clinical trial plans
- Implemented power calculations for sample size planning
- Developed analytics and data visualization for trial outcome reporting, packaging results into R Shiny app
-
Aug 2020 - present
Lead Content Developer and Faculty
SetCONNECT, Inc.
-
Developed materials and content, including assignments, quizzes, projects, and recordings, for several courses
- Data Analytics with Python
- Intro to Statistics and Data Visualization in R
- Business Statistics
- Predictive Modeling for Real-World Problems
-
Developed seminars and introductory courses on statistics for several Indian corporations and educational institutions
- Indo-American Chamber of Commerce
- TVS Group
- Reliance Financial
- ZF CVS
- Vedanta Aluminium
- NITK
-
Worked on several data science and analytics projects for clients
- Warranty cost analysis and predictive modeling for ZF CVS
- Participated in business development strategy and planning, designing core operational tactics for future business directions
-
Developed materials and content, including assignments, quizzes, projects, and recordings, for several courses
-
Oct 2021 - present
Adjunct Faculty
Southern Methodist University
- Taught ‘Doing Data Science,’ an introductory data science course covering the basics of statistics, modeling, and common tech platforms for working professional data science master’s students
- Troubleshooted several issues in the course material related to outdated content
- Assisted on monthly planning meetings to determine scope and design of the data science program
-
Jun 2020 - May 2022
Freelance Consultant
Self-Employed
-
Worked with clients, from academia to industry, on a variety of data analysis and modeling challenges
- Performed analysis estimating genetic content of hybrid plants for Ag Biotech, Inc reducing royalties owed by client by 30%
- Advised PhD researcher on decision tree methods to better model nonlinear effects present in field data
- Assisted several PhD researchers on gathering and analyzing survey data
- Assisted PhD researcher on implementing panel data model for longitudinal data
-
Worked with clients, from academia to industry, on a variety of data analysis and modeling challenges
Academic Experience
-
Apr 2018 - Sept 2018
Research Assistant
University of Texas at Austin - Sandia National Laboratories
- Developed algorithmic routines for consolidating data across experiments into a consensus model, enabling separate modeling of uncertainty within experiments
- Developed benchmark for assessing efficacy of new solutions using previous solution ‘consensus Monte Carlo’
-
Implemented two methods for consensus modeling with a flexibility/scalability tradeoff
- WASP (Wasserstein scalable posteriors) method highly flexible but limited scalability
- Fixed-point iterative barycenter method limited to location-scale families but highly scalable
-
Jun 2016 - Aug 2017
Research Assistant
University of Texas at Austin
- Research assistantship working with Dr. Michael Daniels
- Assisted in research extending ‘network meta-analysis’ methods to scenarios with instrumental variables and missing data
-
Research involved simulation as well as exploratory implementations with the following
- JAGS
- OpenBUGS
- Stan
-
Sept 2012 - May 2019
Teaching Assistant
University of Texas at Austin
- Assisted on statistics courses ranging from introductory to graduate level
-
Ran computer lab sessions on R for students which involved
- presenting and explaining R code
- directing undergraduate teaching assistants
- aiding students with statistical difficulties and computer issues
-
Managed undergraduate teaching assistants for large classes delegating tasks and establishing protocols for
- grading student assignments and quizzes,
- tutoring students
- reporting issues in grading and student performance
- managing and reporting timecards
-
May 2015
Short Course Instructor
University of Texas at Austin
- Taught week-long introductory course on R, developing examples for instruction and balancing basic and advanced concepts for diverse audience needs
- Directed undergraduate assistant instructors, reviewing code and content before each class and setting guidelines for aiding students with statistical and computer related issues
-
Jan 2014 - May 2014
Research Assistant
University of Texas at Austin - Freescale Semiconductors, Inc.
- Worked on product quality analysis, narrowing scope to an unsupervised clustering problem, helping identify systematic product defects
- Implemented spectral dimensionality reduction preprocessing step for more efficient computation
-
Developed unsupervised, latent cluster allocation method and a mixture model for modeling defect patterns
- Implemented a ‘distance-dependent Chinese restaurant process’ for learning product defect patterns
- Used Bayesian hyperparameter estimation for tailoring algorithm performance
- Employed a Gaussian mixture model to model defect patterns on wafer within each identified cluster
- Solution mitigated the need for engineer labor in identifying product defects, saving 100+ hours per product line
Publications
- 1. Ghaffari, N. and Walker, S.G.Parseval’s Identity and Optimal Transport Maps.Statistics and Probability Letters March 2021; 170: 108989
- 2. Ghaffari, N. and Walker, S.G.W2 Barycenters for Radially Related Distributions.Statistics and Probability Letters April 2023; 195: 109788
Honors and Awards
-
2007
- National Merit Special Scholarship (Corporate Sponsor Unilever)
- Walter B. Smith Jr. Undergraduate Scholarship
- University Honors (9 semesters)
-
2010
- Eva Stevenson Woods Endowed Presidential Scholarship
Data Interests
-
Optimal Transportation
- Wasserstein spaces
- Barycenters
-
Heavy-tailed Noise
- Levy-stable distribution
- Levy-jump processes
Other Interests
- Health: adaptogens, exercise science, etc.
- History: ancient civilizations, etc.
- Language: etymology, linguistics, etc.
- Nature: horticulture, weather, etc.