Eric J Ma's Website

I’ve been working on llamabot and wanted to implement a feature to automatically record prompts and LLM responses. I used a PromptRecorder object with a context manager, similar to PyMC’s pm.Model(). However, I faced a challenge: making the context manager aware of its internal code. 🤔 After some research, I found a solution using the built-in contextvars module. By creating a globally-referenceable prompt_recorder variable, I was able to modify the instantiated PromptRecorder object’s state and call its methods while keeping it hidden from the front-end code. 🎉 Now, llamabot can easily record prompt-response pairs! 🤖

🧐I encountered a tricky problem when working on chromatography: I needed to find x-values on concave curves at a specific height without solving simultaneous equations. Check out my blog post 📝 to see how I went from a simple for-loop solution to using JAX's vmap for a more effective, expressive, and compatible solution. This was a fun programming problem to solve! 🔎📈📊😃

🤖 I've been playing with Large Language Models and found some common patterns. 🧠 So, I created a new package called llamabot! 😎 It lets you make your own Python LLM bot, all with just a few lines of code. 🌬️ It should make experimenting a breeze! 🚀 Are you curious? Read on!

Today, I cracked the code on downloading files from Dropbox programmatically! 🎉 Using Python and the Requests library, I can now store large data files for my tutorials on Dropbox, making my Jupyter notebooks portable and internet-friendly. 🌐 It's like having my own personal CDN! 💽

Are you looking to improve your LinkedIn game? 🚀 In this blog post, I explore the best practices and common mistakes when connecting with others on LinkedIn. Learn how to create quality connections, avoid cold connecting without a note, and ensure your profile is up-to-date! 💼 I also share examples of great LinkedIn messages and resources to help you connect like a pro. Don't miss out on these valuable tips to enhance your professional network! 🌐🤝

(Summary was written by GPT-4 and edited by me.)

I built one application using three different GPT libraries 📚, and figured out that it can give me a 100X ROI! 💰 In this blog post, I compare their outputs, developer experiences, and ⏰ discuss how future application developments may look. Curious about the future of GPT and its value proposition? 🚀

(The summary and title were generated by GPT-4 and edited by me.)

Just tried out the Arc browser and I'm loving it! 😍 It's a game-changer for tab management and focus. With features like expiring tabs, tab spaces, and a side-by-side view, it's like having a personalized workspace right in my browser. 🚀 Even has a developer mode for locally hosted sites. Can't wait to see where they take this!

How can a 15th-century explorer teach us about modern goal-setting strategies? In this blog post, I explore Andy Grove's part-historical, part-fictional take on Christopher Columbus in his quest to find a trade route to the Far East. By looking at Columbus' story through the lens of Grove's book "High Output Management," we can learn that OKRs are not a contract but something more useful... Come check out what I learned recently while listening to High Output Management!

I've just automated the creation of Google Docs using Python! 🐍 I used a Google service account to authenticate, then created the doc content using Markdown. I converted the Markdown to HTML, then uploaded it as a Google Doc. I also learned how to set multi-line environment variables and programmatically set document permissions. It was a fun and enlightening project! 🎉

I built a translator using GPT3 for Ark Channel, a daily devotional written in Chinese. The translator parses the text into structured data, translates the devotional and additional texts, and generates cover images.

I also built a UI using Panel and deployed it on a Dokku server on DigitalOcean. The result? My translation time went from 10-15 minutes to just 2! 🚀