Jared Lander
@jaredlander.com
Owner of Lander Analytics Author of R for Everyone Lecturer at Columbia Business School and Princeton University Organizer of NY Open Statistical Programming Meetup jaredlander.com
Working in an air-gapped environment is a good way to discover how much "infrastructure" the internet was quietly handling for you. Actions, images, security tools, updates… without the internet, they become processes involving people, scripts, syncs, and internal repos. Full post: bit.ly/4y8J2TQ
The Air Gap Does Not Remove the Supply Chain
What changes when your CI/CD infrastructure has no route to the internet
bit.ly
We are excited to offer our Agentic Software Development Lifecycle starter kit. Be in touch so we can help your organization modernize their analytic development process.
Lander Analytics@landeranalytics.com · 3w agoYour development platform should make it easier to build software, not create more work. We build standardized, agentic SDLCs with CI/CD, testing, code review, documentation, security scanning, dependency tracking and agent skills built in. See what we can build for your org: bit.ly/46eli4w
Coding agents can write a lot of code quickly. The question is do you still understand the system you're building? Our post looks at using agentic coding without giving up ownership with practical lessons on EDA, ML, schemas and building on patterns your team already knows. Read here: bit.ly/4gHeMbr
You Own the Code Your Agent Wrote
How much of it do you actually need to understand?
bit.ly
We are looking to hire for DevSecOps at @landeranalytics.com. Apply at www.landeranalytics.com/careers/juni...
Junior DevSecOps | Lander Analytics
Explore careers at Lander Analytics, including open roles in data engineering, DevSecOps, data science and more. View current opportunities and apply.
landeranalytics.com
What happens when you change the rules for scoring a competition? Gus Lipkin ran the numbers on competitive bouldering's new scoring system, comparing recent results with the old rules. Most results don't move much but sometimes the differences can be very significant. Read here: bit.ly/3SiaxdZ
Analyzing Competitive Bouldering’s New Scoring System
How the point-based rules change rankings and what they reward
bit.ly
Looking at what has changed in terms of economics and #datascience since I last analyzed the #pizza principle
Lander Analytics@landeranalytics.com · 2mo agoTwelve years after his original Pizza Principle analysis, Jared Lander revisited the data. This time, AI agents helped with the research and coding. The result? A plain NYC slice now costs about 19% more than a subway ride. Read here: bit.ly/4hMmkeW #DataScience #AI #Statistics
Working with text data is completely different from even just a few years ago #databs
Lander Analytics@landeranalytics.com · 2mo agoMost companies have valuable data hiding in comments, PDFs, reports, and spreadsheets. We show how embeddings can turn that unstructured text into something you can search, classify, and analyze using real examples across industries. Read here: bit.ly/4gWUnkl #AI #Embeddings
From thousands of sensor readings to a ranked shortlist. We use ML + pattern matching to identify sites that match key subsurface patterns, turning manual review into a scalable process. 📽️ Watch: youtu.be/saj-VYWvnEw #MachineLearning #Geospatial #DataBS
I've always loved @duckdb.org but this makes it even better #databs
Lander Analytics@landeranalytics.com · 3mo agoIf you've ever hit DuckDB's single-writer limitation, you'll appreciate this. Gus Lipkin takes a look at Quack, DuckDB's new client-server protocol, and why it may finally eliminate the awkward dance of partitioned files, write queues and premature Postgres migrations. bit.ly/4vx8uBi #DuckDB
Everyone wants an AI assistant. Few want company data leaving their network. This walkthrough shows how to build private, self-hosted AI assistant with Hermes, Ollama and DGX Spark—complete with scoped access, audit trails, and persistent memory. bit.ly/4fUEzOF #AI #LLM #AgenticAI #OpenSourceAI
DGX Spark Series (Part 4): Setting up the Hermes Agentic Assistant
Everyone wants an AI assistant. Here's how to build one that keeps your data where it belongs.
bit.ly
The Knicks were given just a 0.4% chance to win Game 4. The Patriots fell to 0.2% during Super Bowl LI. Which comeback was actually more improbable? A look at win probability models, uncertainty, and why "almost impossible" isn't the same as impossible. bit.ly/4ekV43S #SportsAnalytics #NBA #NFL
What Was More Improbable: The Knicks’ Game 4 Comeback or the Patriots’ Super Bowl LI Rally from a 28-3 Deficit?
Using ESPN win probability models, the answer is surprisingly close, and probably closer than the exact decimals suggest.
bit.ly
Do you need enterprise AI infrastructure for a 120M-parameter model? Maybe not-but the answer is more interesting than you'd think. We break down why Chronos-2, DGX Spark, and a simple FastAPI deployment turned out to be the right combination. bit.ly/4eh0FbC #AI #TimeSeries #Forecasting #DGXSpark
DGX Spark Series (Part 3): When the Wrong-Sized GPU Is the Right Call
What we learned serving Chronos-2 from an R-friendly forecasting pipeline
bit.ly
At the New York Open Statistical Programming Meetup on Tuesday, June 16, John Myles White will be there to talk about Genealogy and LLMs! 📍 In-Person at NYU & Virtual | 7 PM (ET) 🙋♀️ More info & RSVP: nyhackr.org #LLMs #DataBS
Use AI coding assistants freely for coding tasks. Use them carefully for analytical ones. In this clip, Jared explains why generated R code still needs statistical judgment, domain knowledge and review discipline before it belongs in a real analysis workflow. Full video here: youtu.be/4csH2-Yf2FQ
The hard part of internal apps usually starts after the prototype works. How do you host it securely? How do users authenticate? How does it scale as more people use it? Jared walks through how Docker, Kubernetes and secure hosting help answer those questions. Full video: youtu.be/7EFYi-aFpdI
The Lander Lens: Fast, Secure & Scalable Internal Apps (Part 2)
YouTube video by Lander Analytics
youtu.be
AI can build a working app surprisingly fast. That does not mean it is ready for production. In our latest blog post, Jared Lander breaks down where AI coding agents help, where they fall short and why human judgment still matters in software engineering. bit.ly/3PQn0V5 #AI #Engineering #LLMs
Vibe Coding Gets You to a Working Demo, But Engineering Is What Makes It Hold Up
AI coding agents are becoming useful interns, but teams still need judgment, taste, infrastructure and review before the code is something they can actually run.
bit.ly
TONIGHT is the May New York Open Statistical Programming Meetup! Christian Martinez dives into NYC Open Data. Make sure to RSVP by 3 PM (ET)! 📆 Thursday, May 14 | 7 PM (ET) 📍 In-person at NYU & Virtual 🙋♀️ More info & RSVP: nyhackr.org #NYCOpenData #DataBS
Your coding agents probably don’t need RAG yet. At Lander Analytics, we connected 16+ repos using shared markdown files, grep, and zero new infrastructure to give agents cross-repo context that actually stays current. bit.ly/42ulRFy #AI #LLM #CodingAgents
You Probably Don’t Need RAG for Your Agent’s Knowledge Base
We connected 16 repos with markdown files, grep, and zero new infrastructure
bit.ly
Traditional #machinelearning is still a great tool #databs
Lander Analytics@landeranalytics.com · 5mo agoWith LLMs and pretrained models everywhere, it’s tempting to skip traditional modeling. But train-test-evaluate still wins when domain knowledge matters. This post breaks down when old-school workflows outperform the shortcuts. bit.ly/42OtuGQ #DataScience #MachineLearning #AI
From raw customer behavior to real-time churn risk. See how predictive models and dashboards help teams spot risk early and take action before customers leave. Watch the video: youtu.be/D8xbXS0rsI0 #PredictiveAnalytics #CustomerChurn #MachineLearning #DataBS
LLM tool calling began as a workaround for stale data and evolved into the backbone of agentic AI. We walk through the mechanics, the evolution from XML to native APIs, and why this simple idea unlocked powerful workflows. Read more: www.landeranalytics.com/post/the-ana... #AI #LLMs #AgenticAI
The Anatomy of an LLM Tool Call
How a prompt engineering trick unlocked agentic AI
landeranalytics.com
Good memories working with Mike Band on the Vikings draft
Lander Analytics@landeranalytics.com · 5mo agoWith the NFL Draft this week, we’re revisiting how we helped the Minnesota Vikings build a draft model in 2015. From messy scouting data to “blue sticker” picks, this is how analytics made it into the war room & helped shape an elite class. 📝: bit.ly/4u2AQ5j #NFLDraft #SportsAnalytics #BigData
A better cron
Lander Analytics@landeranalytics.com · 6mo agoCron works, until it doesn’t. No logs, no alerts, no visibility. This post breaks down where cron falls short, when it still makes sense, and how tools like rpeat make scheduling jobs more reliable and debuggable. bit.ly/4cbcCQk #DataEngineering #DevOps #OpenSource
Check out how we power through large geospatial data
Lander Analytics@landeranalytics.com · 6mo agoRendering millions of geospatial points used to break browsers. Not anymore. This post walks through the modern stack, from Leaflet to deck.gl + Arrow, and how teams scale interactive maps from thousands to millions of points. bit.ly/41lHZkV #Geospatial #DataViz #OpenSource
We're using LLMs to do more with more
Lander Analytics@landeranalytics.com · 6mo agoAI adoption isn’t about replacing people, it’s about enabling them. Our blog post explores why most companies aren’t ready for AI-driven layoffs, where the real value comes from, and how to use AI to upskill teams and build better workflows. bit.ly/414YRMI #AI #FutureOfWork #Leadership
Loving the #geospatial #databs
Lander Analytics@landeranalytics.com · 7mo agoGeospatial datasets keep growing—and file format choices can quietly inflate storage and slow every downstream step. In this post we compare common formats, explain why GeoParquet is gaining traction, and when it’s time to move from files to PostGIS. bit.ly/3PsGxud #Geospatial #DataBS
Some thoughts about open source vs closed models from Joe Marlo in the age of #GenAI
Lander Analytics@landeranalytics.com · 7mo agoOpen source LLMs are closing the gap fast. This post looks at when they’re “good enough” for production, when closed models still make sense, and what the real costs of self-hosting look like beyond API pricing. Full breakdown here: bit.ly/4u3y1S7 #OpenSource #AIinPractice #DataScience
Some first thoughts on using @anthropic.com's Claude Code with #rstats
Lander Analytics@landeranalytics.com · 7mo agoWhere do AI coding tools actually help R users, and where do they quietly fall short? Our blog post digs into Claude Code as an R assistant: great for syntax, Shiny, and infra work, risky for analytical judgment. Based on real use at Lander Analytics. 🔗 bit.ly/3MWnMyt #RStats #OpenSource #DataBS
