Level 1 Beginner
Architecture
HTMLprompting perplexity to assist with architecture and construction runoffs
Who is this for? construction runoff notes
Sam Castillo · github.io
README
Sam Castillo is a Massachusetts entrepreneur. He studied mathematics and actuarial science at the University of Massachusetts Amherst, climbed through corporate analytics, and now owns Predictive Insights AI and LEARN AI.
Clients hire Sam and buy the products at Predictive Insights AI. SamWiki is the proof behind that business: live demos, source, and notes in one index, ranked so a customer, a student, or a collaborator can see what shipped and how hard it is.
There is always someone smarter out there. The edge is putting the work where people can use it together. This is not a zero-sum game. It is a cooperative one.
Sam
Trust
Notes from clients, collaborators, and a mentor.
He was able to accomplish this all under the budget I set aside for the project. He discovered patterns in the data that I never expected to see, pairing his expertise with my industry knowledge perfectly. I’d strongly recommend Sam for any data task you may have.
Besides being easy to work with, he has been extremely consistent with project management. We have had many tight deadlines and last-minute changes, but he always provided me with clear instructions. His talent, demeanor, professionalism, and attitude speak volumes.
Sam took action and created a data visualization survey to better understand which tools others were using so that we could tailor design decisions to match our end users. He goes out of his way to improve the work experience for others. He takes action when it is needed.
Sam is making predictive analytics approachable and accessible to new talent in the actuarial field. It’s a tough subject and a hard thing to do, but he has proven he’s up to the task. Sam is the kind of hard working, humble, smart person everyone wants to work with.
He immediately stood out as someone that values hands-on experience over theorizing. Sam’s real world experience on how to promote a business would make him an asset to any team.
Technical projects can often have many unknowns, but Sam always knew and expressed what the overarching end goal was. The coordination he did between the different members, and making sure everyone had the things they needed, was the reason everything moved along.
Difficulty is a level, the same way a game asks you to pick one before you play. Choose a rank and test yourself. The line on each card says who the project is for.
Every card has Contribute / Open a PR. Fork the project, make the improvement, and send it back. You are joining the team.
SamWiki lists the original public work in mathematics, machine learning, and actuarial science. A Live link opens a demo. Source opens the repository. Notes stay next to the code.
Much of the math was built at the University of Massachusetts Amherst, on Sam’s own time. The pages are being brought up to one GitHub Pages look.
Search by name, language, topic, or who a project is for. Forks of other projects stay off this index. Training notes from the previous homepage are on the Running page. Also on this site: Home, About, and Code.
For a forecast, a model, or a product, go to Predictive Insights AI. Linktree holds the other off-site links.
No public projects match that search.
Notes, sites, and first looks. A place to start, then pass what you learn back to the team.
Level 1 Beginner
prompting perplexity to assist with architecture and construction runoffs
Who is this for? construction runoff notes
Level 1 Beginner
—
Who is this for? nonprofit site visitors
Level 1 Beginner
images for website to conserve space
Who is this for? site asset browsing
Level 1 Beginner
University of Massachusetts Amherst 2017, Stats 525
Who is this for? first stats projects
Level 1 Beginner
PredictiveAnalyst site
Who is this for? browsing an older site
Level 1 Beginner
for security data and AI alarm access control systems
Who is this for? security data sketches
Level 1 Beginner
Ideas about helping people get treatment for less money
Who is this for? health-cost idea sketches
Level 1 Beginner
Companion to the Text Introduction to Statistical Learning
Who is this for? ISL reading companions
Level 1 Beginner
Materials for website before launch
Who is this for? old landing-page files
Shiny apps and guided tutorials. The central limit theorem explorer is this level.
Level 2 Intermediate
MUSIC
Who is this for? musicians who code
Level 2 Intermediate
Useful R tutorials for Actuaries
Who is this for? actuaries moving GLMs to Excel
Level 2 Intermediate
Explanation, discussion, simulations, and comparisons of clustering methods
Who is this for? first clustering notes
Level 2 Intermediate
Slides for my presentation at the 2019 SOA Predictive Analytics and Futurism Symposium
Who is this for? analysts learning stories
Level 2 Intermediate
Competition to predict home prices
Who is this for? first GBM notebooks
Level 2 Intermediate
Movie and actor comparison by cosine similarity and euclidean distance measures
Who is this for? first vector models
Level 2 Intermediate
A Shiny App of the CLT
Who is this for? first Shiny apps
Level 2 Intermediate
A container for data sets to help actuaries who are practicing predictive analytics
Who is this for? actuarial exam prep
Level 2 Intermediate
An open-source set of study notes using Kaggle
Who is this for? actuarial exam prep
Level 2 Intermediate
Scilab linear algebra example of R code
Who is this for? linear algebra students
Level 2 Intermediate
Plotting a network of recommended movies from imdb.com.
Who is this for? first network graphs
Level 2 Intermediate
Local Text To Speech Model
Who is this for? first local TTS scripts
Level 2 Intermediate
A collection of textbooks, research papers, and tutorials
Who is this for? self-study readers
Level 2 Intermediate
—
Who is this for? crash-data explorers
Actuarial depth and models you finish yourself. Loss development triangles are this level.
Level 3 Advanced
Source code from a project from with BCBSMA this summer
Who is this for? health analytics teams
Level 3 Advanced
ML tutorials and Shiny app
Who is this for? actuarial exam prep
Level 3 Advanced
An online study guide for the SOA's predictive analytics exam.
Who is this for? actuarial exam prep
Level 3 Advanced
A Shiny App for detecting abnormal stock price fluctuations
Who is this for? time-series learners
Level 3 Advanced
Bitcoin price anomaly detection using R and Shiny
Who is this for? time-series learners
Level 3 Advanced
Dolphin Mistral FastAPI hybrid app with Grok integration, Supabase, and more.
Who is this for? hybrid LLM app builders
Level 3 Advanced
—
Who is this for? interval modelers
Level 3 Advanced
A submission to a kaggle competition
Who is this for? regression competitors
Level 3 Advanced
Study Notes and Code from Modern Actuarial Statistics I
Who is this for? actuarial exam prep
Level 3 Advanced
Field notes from a Hetzner + Tailscale + Nginx + Ollama build: how to run a local LLM ethically and safely for nonprofit workforce training.
Who is this for? self-hosted LLM operators
Level 3 Advanced
Grok Bot + Dolphin-Mistral hybrid (IMDb multi-level FastAPI prompts)
Who is this for? local LLM builders
Level 3 Advanced
Code and data files for the article "Painlessly Merge Data into Actuarial Loss Development Triangles with R"
Who is this for? reserving analysts
Level 3 Advanced
—
Who is this for? exam-data modelers
Research math and machine learning. Prior weights in XGBoost, the Riemann zeta function, and P vs NP live here.
Level 4 Painfully challenging
A method of adding prior weights to xgboost
Who is this for? ML practitioners
Level 4 Painfully challenging
—
Who is this for? complexity theory readers
Level 4 Painfully challenging
—
Who is this for? analytic number theory
Level 4 Painfully challenging
Website for math, using AI to prove math theorems
Who is this for? AI-assisted math