Mon
- Coffee chat and pairing with @Elizabeth on today’s Leetle
- Finally fixed the user flow for saving letters in Open When, ready for tomorrow’s demo-palooza. It’s a hacky fix but it’s good enough for me.
- Started going through the Django tutorial to implement my own backend to Open When
- Read about ways that people have trained AI agents to play Mancala. Interested in exploring that post-batch, among other things!
Tues
yesterday:
- Mushroom image classification ML Jam – thanks for organizing @Keerthi
- Presented Open When at demo-palooza and got to see other people’s demos – really liked seeing @Teon’s scrapbooking site and @Basil’s ray tracing maze game!
Weds
- Got my training and validation DataLoaders working for my mushroom image classifier on Colab. I realized it’s been about a year since I took my ML courses so it’s nice to refresh myself while looking at pictures of mushrooms. I think there’s some interesting ways I can take the project like trying a compositional/hierarchical approach, interpretable/explainable architectures (attention mechanisms??), transfer learning, fine-tuning, heck maybe generative models like transformers and diffusion.
- Chatted with @Suhith about ML infra, ML performance, and the self-driving industry
- Caught up with @Kevin and did a nice DP problem
- Non-programming talks: I want to read my height in books, would be easier for me than @Spencer haha and I’m glad I was reminded of Neetcode and other great job search resources thanks to @Elijah for future SWE internships (or full-time jobs, if it turns out I get rejected from all my grad schools)
- Gave my last non-programming talk in batch, Staying in my Lane! Thank you to @Suhith and @Kevin for encouraging me to present it — I was so close to not doing it because I was like, I don’t even have anything together really… but I got really nice feedback on it and I’m really happy I did it
- Did today’s Leetle (with help from chatgpt) and yesterday’s Leetle (I wrote my own solution inspired by Kadane’s algorithm for finding the max subarray sum, which I’m proud of because I could never remember the details to Kadane’s algorithm!)
Today was a good day.
edit: so good that I forgot to mention I got rejected from CMU Robotics lol. and i simply cannot be bothered to feel bad about it
Thurs
- Popped in on Fast.ai group! Really enjoyed the meeting (thank you for presenting @Karen ) and planning on continuing to come next month
- Paired with @Karen on setting up Tailscale and nginx on my old linux MacBookPro so I can self-host my apps and who knows, maybe I’ll store photos and docs and stuff too. This was so so fun and I really felt the magic of RC
- Presentations and niceties and end of batch celebration
edit: thank you @Sam for being with me as I opened my rejection letter from UW
Fri
today:
- End of batch reflection — planning to write it up into a return statement this weekend
- Backed up my college files on my linux Mac using Tailscale taildrop – loving this new tool, thanks @Karen
- Chatted/paired with @Nadja — I need to try running Ollama locally on my M2 Mac!
- Had the honor of signing @Ross ’s copy of my book The Glitch (signed @Jess ’s the other day, and thanks to @Marc and @Priya for buying copies too)
- Finally finished writing niceties throughout the day.
What a joy it has been. Have moved to #alumni checkins>Grace Kwak for future updates.