Finishing the course | Readings
Finished the Udemy course! Took me just getting focused and setting a plan (obviously). For the next week, I’ll be on vacations but as I’ve gotten my talk accepted for the Volkswagen Tech summit, I’ll start working on it (16 of May) before going to London. Going to be talking about using SQL to empower our decisons and be pushing for Duckdb and other tools like steampipe on the presentation. Regarding the exam I’m feeling that it’s quite challenging from the practice test but I got a 70% on my first try and had time enough to review it. So I’ll probably schedule it for the month of June.
I’ve also gotten a personal project that I’ll keep the wraps on it for now. It’s related to getting a RAG app and I’d love to use llama33 or another open model for it. As such, I’ll probably test using ollama and get it running on ASS fargate or something similar. But my goal will be to open source it if it works well.
Readings of the week
I haven’t been able to get a lot of reading and failed to read the AWS glue documentation. However here are the following articles I’ve read in the past 2 weeks:
- Llamafile’s progress, four months in - Mozilla Hacks - the Web developer blog by Stephen Hood: Really interesting to see both llama.cpp and llamafile improvements. Been using only ollama locally and still haven’t found a use case to use either of the projects directly
- AWS re:Invent 2023 - Accelerate generative AI application development with Amazon Bedrock (AIM337) by AWS Events: Taken an overview look on the bedrock service. Might be testing it in the next couple of months for deploying an agent
- How We Built Slack AI To Be Secure and Private - Slack Engineering by Kelly Moran: The usage of sagemaker for the rag is a good validation that the tool although expensive can be very useful for deploying rag apps
- Cheaper, Better, Faster, Stronger | Mistral AI | Frontier AI in your hands by Mistral AI: Really good to see mistral deploy a better model
- How to succeed as a data engineer without the burnout - Stack Overflow by Katja Belova: Small blog talking a bit about a field that requires a lot of knowledge on a lot of systems. At this point I’m thinking they are the full stack of the data world
- Trying out Rye: Second time hearing about rye. As a user of poetry this also looks like an interesting tool to use as it uses the native venv but provides the whole tool suite like rust’s cargo
- Spark 3 and GPUs to Reduce Cloud cost by up to 70% | The PayPal Technology Blog by Ilay Chen
Goals for this week
- Write the presentation for the tech summit
- Schedule the exam
- Start the personal project