S T O R Y D E S K

Financial Services

---

• Client: Valuation and corporate finance advisory firm
• Date: June 2008 - Aug. 2008
• Dept: Information Technology
• Business service: Data analysis
• Technical service: Developed a key performance indicators (KPI) tool using data from an Interaction database.
• Technologies: SQL

• Client: U.S. automotive warranty financing firm
• Date: Oct. 2007 - May 2008
• Dept: Information Technology
• Business service: Payment processing, PCI compliance, fraud detection
• Technical service: General application development and database management.
• Technologies: SQL, C#, .NET, WinForms

• Client: Global bank and credit card firm
• Date: June 2005 - Sep. 2007
• Dept: Business systems
• Business service: LAN systems engineering
• Technical service: Directory services development and directory design.
• Technologies: C++, C#, .NET, Active Directory

Trading, Investing

---

• Client: Global capital markets technology firm
• Date: Oct. 2017 - Apr 2023
• Dept: Engineering
• Business service: Trade surveillance product development
• Technical service: Trade surveillance model dev lead.
• Technologies: Java, Python, AWS {EC2, S3, Firehose, Redshift, DynamoDB}, H2O, scikit-learn, Spark, JavaScript, git, regex, SQL, Jenkins (CI/CD), Jetty

• Client: Nationally recognized investment bank
• Date: Jan-July 2010
• Dept: Information Technology
• Business service: Trading systems
• Technical service: Member of a team of consultants and internal employees developing a mission-critical, client-facing, bond trading web application.
• Technologies: C#, ASP.NET, JavaScript

• Client: Chicago trading software firm
• Date: Jan. 2009
• Dept: Information Technology
• Business service: Web Development
• Technical service: Worked on a team to developed a prototype web application to receive and display a stream of XML-formatted trade data incoming at a rate of 200 trades/sec.
• Technologies: C#

Side Projects

---

• AI/LLM engineering:
  • Automated LLM prompting for text summarization.
  • Expert-routing: Built a LLM-driven router & trained 2 PEFT adapter models to use atop a base LLM for topic Q&A.
  • Quantized the Mixtral 8x7b LLM via MLX.
  • Implemented from scratch Vaswani et al's Transformers architecture from "Attention is All You Need".
• Trading:
  • Wrote & backtested proprietary equity/ETF, futures, & crypto trading strategies in vectorbt.
  • Monte Carlo simulation for financial planning.
  • Daily returns forecasting and probability analysis.
• ML in medicine: Heart disease classification in PyTorch, Spark + MLib, & sklearn.
• Descriptive stats in K-12 education: US Per-State Public Education Value-for-Money Comparison

Health & Medicine

---

• Client: Physiologic
• Date: May 2016 - June 2016
• Dept: Firmwide
• Business service: Technical co-founder (pre-seed startup); data science
• Technical service: Created multiple classification models (machine learning) of a user of a health monitoring device. Advised on technical and statistical issues.
• Technologies: scikit-learn, Python

• Client: Varsa Health
• Date: Oct. 2014 - Dec. 2014
• Dept: Firmwide
• Business service: Technical advisory
• Technical service: Advised on initial technical architecture and created an initial cloud-based data infrastructure and customer-facing data flow.
• Technologies: NoSQL

• Client: Chicago area hospital
• Date: Nov. 2009 - Jan. 2010
• Dept: Payment Processing
• Business service: Efficiency improvement
• Technical service: Worked on a team to deliver an application that reduced payment processing workload by over 95%. Developed a custom rules engine at the core of this processing.
• Technologies: C#, .NET, WinForms

Research: Medical Imaging & Mental Health

---

• Client: DePaul University - MedIX Lab
• Date: Dec. 2014 - June 2016
• Dept: Computer Science
• Business service: Imaging in service of lung cancer research and bioscience research
• Technical service:
  • Developed an image processing system to process LIDC images and produce contoured imagery containing cancerous nodule contours.
  • Built in TensorFlow, trained, and achieved better-than-random first-shot accuracy a deep convolutional neural network for LIDC lung cancer nodule classification.
  • Developed a generalizable web based data visualization tool for time series motion data on a project involving video tracking of the c.elegans nematode.
• Technologies: TensorFlow, pandas, Python, MATLAB, JavaScript, plotly

• Client: Northwestern University - Center for Behavioral Intervention Technologies
• Date: Mar. 2012 - Aug. 2014
• Dept: Feinberg School of Medicine
• Business service: Full stack, full lifecycle software development
• Technical service: Responsible for developing web and Android apps, mobile data collection systems, leading junior staff, consulting for multiple external collaborators, and administering multiple servers in-house. Publicly visible and shareable products include:
  • Metamorphoo -- A Node.js-based data collection system through which over 1TB of data was collected from in-field mobile devices and stored in a Postgres database in 1.5 years.
  • My Mantra -- An app to help motivate and inspire users to be their best selves.
• Technologies: NoSQL, SQL, MongoDB, Postgres, JavaScript, node.js, Java, Android development

Retail

---

• Client: Leading art supplies firm
• Date: Aug. 2010 - Mar. 2012
• Dept: Marketing
• Business service: Pricing
• Technical service: Close collaboration with a pricing department to develop and apply pricing algorithms in the catalog and flyer channel. Developed an internal web application to enable multiple teams to collaboratively price products.
• Technologies: SQL, SQL Server, C#, .NET, ASP.NET, JavaScript

Manufacturing

---

• Client: Global manufacturing firm
• Date: Feb. 2009 - Sept. 2009
• Dept: Information Services
• Business service: Web Development
• Technical service: Led a team of developers to create and deliver a public-facing website built on a content management system.
• Technologies: C#, ASP.NET, JavaScript, CSS

Professional and Media Services

---

• Client: Inventum Digital, Inc.
• Date: May 2016 - July 2016
• Dept: Firmwide
• Business service: Outsourced CTO and software architecture service
• Technical service: Working on a digital marketing reboot for a downstream client involving REST service architecture, UI/UX redesign, data visualization, and collaboration with overseas developers.
• Technologies: REST, JavaScript

• Client: Photography firm
• Date: Sept. 2013
• Dept: Information Technology
• Business service: Reporting
• Technical service: Advised, implemented, and deployed a Microsoft SQL Server Reporting Services deployment.
• Technologies: SQL, SSRS

• Client: National presentation services firm
• Date: June 2008 - Dec. 2008
• Dept: Information Technology
• Business service: Web Development
• Technical service: Worked on a team to create and deliver an internal image asset management website.
• Technologies: C#, .NET, ASP.NET

• Client: International management consulting firm
• Date: June 2008 - July 2010
• Dept: Integration
• Business service: Custom application development and data management
• Technical service: Technical consulting for various firms across several industries via this full service consultancy.
• Technologies: SQL, JavaScript, CSS, C#, .NET, ASP.NET, WinForms, Java

Non Profit

---

• Client: Nonprofit engineering firm
• Date: Jan. 2009 - Feb. 2009
• Dept: Information Technology
• Business service: Request for Quotation (RFQ)
• Technical service: Worked on a team to build a streamlined, web-based replacement RFQ process.
• Technologies: C#, .NET, ASP.NET

• Client: International religious client
• Date: Various 2008-2010
• Dept: Information Technology
• Business service: Web Development
• Technical service: Some web development for a religious client.
• Technologies: SQL, C#, .NET, ASP.NET

Artificial Intelligence, Large Language Models

---

AI services:

• Prompt Engineering - "shape input to get good AI output"
• Retrieval Augmented Generation (RAG) - "ground my AI's responses in my documents"
• LLM fine-tuning (PEFT) - "update the LLM to focus on my task"
• Agentic AI - "let AI perform a task autonomously"

Machine Learning Engineering, Data Science, Statistics

---

ML & data science services include:

Machine Learning

• Regression - "predict number"
• Classification - "predict type"
• Clustering - "predict similarity"
• Recommendation - "predict match"
• Social network analysis - "describe or predict entity (who/what/when/where) relations"
• Deep learning (e.g. multilayer convolutional neural networks) - "model complex data"



Imaging, Vision

• Image processing - "transform image/video"
• Image analysis - "generate features about image/video"
• Object recognition - "identify objects in this image/video"



Natural Language Processing

• Text mining - "extract signal from text"
• Sentiment analysis - "what feeling is in the text?"



Statistics

• Predictive modeling with a focus on model interpretability - in contrast to an ML approach, which seeks to maximize some performance metric (F-beta, R^2, etc.)
• Hypothesis testing (frequentist or Bayesian) - "if I change X, how and by how much does Y change?"
• Descriptive statistics - summarize the data



Data Engineering

• Data fusion - "join/relate/combine datasets"
• Data cleaning - "prep data for modeling"
• Database design and data modeling - "organize and structure the data"
• Reporting - "generate reports"

Fractional CTO/CDO,
Team Leadership

---

Technical Strategy and Team Leadership for:

• AI/LLM engineering
• ML engineering & data science
• Cloud - AWS
• Back-end software & data engineering
• Web and mobile app development
• Testing (unit, integration, performance)

Management Consulting

---

Management consulting services:

• Startup advising
• Product/service ideation
• Product advisory, analysis, and decision support
• Product management
• Project management
• Pricing strategy and modeling
• Business process analysis
• Team analysis