Masoud Masoumi
banner
masoudmim.bsky.social
Masoud Masoumi
@masoudmim.bsky.social
Engineer turned Data Scientist | Interested in history | All views are personal | Personal website: masoudmim.github.io
Reposted by Masoud Masoumi
Check out this blog post from @clairemkbowen.bsky.social showing how federal data shape your entire day—often without you even noticing. Federal data are everywhere! It’s the invisible infrastructure powering daily life and the big decisions that shape our futures. #statssky
A Day in the Life with Federal Government Data – Association of Public Data Users
apdu.org
November 26, 2025 at 11:02 PM
Reposted by Masoud Masoumi
Olmo 3 is notable as a "fully open" LLM - all of the training data is published, plus complete details on how the training process was run. I tried out the 32B thinking model and the 7B instruct models, + thoughts on why transparent training data is so important simonwillison.net/2025/Nov/22/...
Olmo 3 is a fully open LLM
Olmo is the LLM series from Ai2—the Allen institute for AI. Unlike most open weight models these are notable for including the full training data, training process and checkpoints along …
simonwillison.net
November 23, 2025 at 12:17 AM
A few photos from my trip to Maine in September.
November 19, 2025 at 1:01 PM
Reposted by Masoud Masoumi
Hooray for an introvert!
November 16, 2025 at 10:42 PM
finally caught Waiting for Godot on Broadway. it was a good show! I've got great respect for Keanu Reeves, both as an actor and as a person.

#WaitingForGodot #ActorAppreciation #Broadway
November 10, 2025 at 3:24 PM
Reposted by Masoud Masoumi
Billy Joel releases his “Piano Man” LP this week in 1973.

“I was shocked and embarrassed when it became a hit,” he said of the title track. “The melody is not very good .. the lyrics are like limericks. .. But my songs are like my kids and I look at that song and think, ‘My kid did pretty well.’”
November 9, 2025 at 6:36 PM
I started using the ThunderAI add-on in Thunderbird. Now my local LLM automatically classifies, auto-replies (reviewable), and summarizes emails across multiple accounts. I did not expect to appreciate it this much!

#Thunderbird #ThunderAI #Productivity #PrivacyFirst #LLM
November 7, 2025 at 2:59 PM
I wrote a short blog post about the idea of how two evolutionary cognitive abilities can help educators and students create more effective teaching-learning relationships.

#Education #Teaching #Learning #HigherEducation #Pedagogy

masoudmim.github.io/blog/2025/ev...
October 27, 2025 at 12:26 PM
when someone asks me why I'm coding and doing data science work
October 3, 2025 at 9:31 PM
Reposted by Masoud Masoumi
Once upon a time in Niagra Falls
August 3, 2025 at 2:49 PM
I had been meaning to write this piece, which I would call my statistically supported argument
- for more funding for research
- against funding mainly successful researchers, and
- against trying to optimize research funding allocation

#ResearchFunding

masoudmim.github.io/blog/2025/di...
Why More Beats Best | Masoud Masoumi
A statistical argument against supporting mainly successful researchers and optimizing research allocation
masoudmim.github.io
July 2, 2025 at 10:03 PM
Reposted by Masoud Masoumi
Explore Wikipedia through a data map. Pages are grouped by semantic similarity, for topic clusters.
Hover to see details, zoom to explore more fine-grained topics, click to go to a page. Search by page
name to find interesting starting points for exploration.

lmcinnes.github.io/datamapplot_...
June 22, 2025 at 3:36 PM
I wrote a simple RAG-based procedure as an example for reviewing the procedure and providing a quick and interesting way of learning RAG.

It walks you through the development of a vector database, and then a simple application via Ollama, Milvus, and Streamlit.
masoudmim.github.io/blog/2025/ra...
June 22, 2025 at 10:17 PM
Reposted by Masoud Masoumi
Our computer vision textbook is now available for free online here:
visionbook.mit.edu

We are working on adding some interactive components like search and (beta) integration with LLMs.

Hope this is useful and feel free to submit Github issues to help us improve the text!
Foundations of Computer Vision
The print version was published by
visionbook.mit.edu
June 15, 2025 at 3:45 PM
I trained a logistic regression model on the source data and then evaluated its performance on both the source and target domains to measure the performance degradation caused by the covariate shift.

The production performance slowly degrades because the feature relationships changed.
June 6, 2025 at 11:09 PM
Wrote a post outlining the step-by-step process of implementing a PINN for a simple one-dimensional heat transfer problem. I hope this approach makes the topic more accessible to undergraduate students and provides a clearer understanding of how they work.

masoudmim.github.io/blog/2025/pi...
May 31, 2025 at 12:35 PM
About five years ago, I began teaching Python programming to undergraduate engineering students for the purpose of data analysis. 1/6
May 20, 2025 at 11:55 PM
"Reading for pleasure has plummeted over the past 20 years"
... respondents who read for pleasure on any given day declined by an average of 2 per cent every year from 2003 to 2023.
archive.ph/ONrf6
archive.ph
May 5, 2025 at 10:18 AM
I just published a post on converting text into a vector database using Milvus. I have found Milvus to be a great tool for NLP projects.
You can check out the post here: masoudmim.github.io/blog/2025/te..., and the associated code on GitHub here: github.com/MasoudMiM/te...
github.com
April 14, 2025 at 12:07 PM
Reposted by Masoud Masoumi
Meta just dropped Llama 4 on a weekend! Two new open weight models (Scout and Maverick) and a preview of a model called Behemoth - Scout has a 10 million token context

Best information right now appears to be this blog post: ai.meta.com/blog/llama-4...
The Llama 4 herd: The beginning of a new era of natively multimodal AI innovation
We’re introducing Llama 4 Scout and Llama 4 Maverick, the first open-weight natively multimodal models with unprecedented context support and our first built using a mixture-of-experts (MoE) architect...
ai.meta.com
April 5, 2025 at 7:53 PM
As part of my "Data-Driven Problem Solving" course for engineering students, I do a review of Python programming. I put the videos for that section of the course on YouTube in case someone else finds them useful: youtube.com/playlist?lis...
Python Overview - YouTube
This series of short videos is designed for the "Data-Driven Problem Solving" course, in which I review the fundamentals of Python programming.
youtube.com
April 2, 2025 at 12:58 PM
Reposted by Masoud Masoumi
Part 2 of SLAM handbook is out for public comments! let us know what you think :-) Issue tracker on GitHub awaits! Link: github.com/SLAM-Handboo...
March 31, 2025 at 8:44 PM
What if you could implement a Large Language Model (LLM) that combines stock price data, companies' financial reports, and public news to conduct financial market analysis?
That's exactly what I explored in this project: github.com/MasoudMiM/st...
There is a lot of room for improvement.
GitHub - MasoudMiM/stock-news-analysis: The Stock Analysis and Recommendation System is a tool that provides investors with actionable insights by analyzing financial data and news. It integrates APIs...
The Stock Analysis and Recommendation System is a tool that provides investors with actionable insights by analyzing financial data and news. It integrates APIs for data retrieval, uses machine lea...
github.com
March 9, 2025 at 8:34 PM
A free personal tool that you can run yourself to pull stock prices, conduct performance analysis, and email you a summary report of the best and worst-performing stocks.
I used to run it on my local server daily, receiving summary reports every morning.

#StockMarket #Investing #InvestmentTools
GitHub - MasoudMiM/stock-performance-analyzer: A Python tool for analyzing stock performance over a specified period. It fetches data from Yahoo Finance, calculates key metrics like price growth and volatility, and identifies top and bottom performers. The tool sends email reports with analysis results and attachments using the Gmail API, making it ideal for investors and analysts.
A Python tool for analyzing stock performance over a specified period. It fetches data from Yahoo Finance, calculates key metrics like price growth and volatility, and identifies top and bottom per...
github.com
March 6, 2025 at 3:35 AM