|
|
|
|
MLOps crew, are you ready for Day Two? Another packed day of AI in
Production awaits you!
Hmmm.. could be the start of another song there...
While I write that, we'd greatly appreciate it if you could grant us a moment of your time to complete our evaluation survey.
And speaking of grants... Prem launched a Gen AI Grant Program, offering open-source model API
usage and free fine-tuning jobs for both closed and open-source models. Find out more and apply here.
|
|
|
|
|
|
|
Data
Governance and AI // Alexandra Diem // MLOps Podcast #212
After chatting with Alexandra, I imagined her to be like the head of an F1 pit crew – lightning-fast, and laser-focused to keep the data engine motoring.
We start by talking about her academic journey, including disproving a theory about Alzheimer’s, and how she transitioned to the world of machine learning operations. She talks about her team's approach to collaboration and software best practices, tackling the challenges of AI in sensitive environments and prioritizing use cases through agile methods.
We also talk about the shift towards self-serve data solutions and AI-powered workflows, highlighting the crucial role of cloud data platforms and the software engineering skills essential for data scientists.
Buckle up for a great podcast!
|
|
|
How to Create High-Performing Gen AI Applications - with TruEra
|
|
|
|
|
|
|
Developers are having a field day creating new AI applications. But the vast majority never get out of the lab and into production. Are reliable Gen AI apps just a hallucination?
Hear the latest on production-grade Gen AI apps from experts at OpenAI, Microsoft, TruEra, and Wing Venture Capital. On February 28th, this webinar will cover:
• Beyond LLMs – how multimodal is taking hold
• What does testing and evaluation even mean?
• How to think about the tech stack, tools, and even foundation models in an era of rapid innovation
Register here
|
|
|
|
|
|
|
Evaluating and Integrating ML Models // Morgan McGuire and Anish Shah // MLOps Podcast #213
It'd be a loss to the ML world, but a little bit of me is hoping Morgan and Anish go back to their university days. Morgan, so he can get his
radio gig back and promote my latest banger, and Anish to help with the laser show.
I don’t think they will though, as they love what they’re doing, which is a lot: in-person events, blog writing, building integrations, developing courses, and more. But, they especially enjoy interacting with the community to find out what they’re building so they can share it, like the Allen Institute’s OLMo and the Open RL Benchmark by CleanRL.
We got talking a little bit about how they view LLMs internally, and a lot about evaluation! And it seemed like Morgan did slip back into his radio hosting days, asking me a great question along the lines of, if a model gives an answer that gets the right result but isn't the preferred way to do it, is that right, or wrong?
I now feel I might have to go back to university and study the philosophy of right and wrong.
|
|
|
💡Job of the weekFounding Software Engineer // Cleric (US, San Francisco) Cleric is banishing humans from the production environment. They’ve started with an autonomous AI agent that frees software engineers up from on-call support. Cleric is VC backed and looking for a maniacal builder who wants to get in early.
Responsibilities:
- Craft the core components of the agent, including memory, planning, and reasoning.
- Own product decisions to Iterate toward product-market-fit.
- Ship a high-scale agent capable of concurrently managing thousands of services.
- Give a damn about quality.
Requirements:
- You are fluent in Rust, C++, Scala or Go.
- Strong software engineering
fundamentals.
- Deep experience with on-call & sprawling production environments.
- You are obsessed with generative AI.
|
|
|
|
|
|
|
Getting your dad to teach a lesson isn't unusual, it's Pa for the course! Boom boom! Much better than that joke, and created my MLOps folks in the know and not my dad, are these two courses:
|
|
|
MLOps Community IRL Meetup
|
|
|
|
|
Data Streaming in Action: From Kafka to Flink // IRL Meetup #65 Stockholm
Some interesting questions to start this one, like what's your favorite DS/DE editor, and what the biggest achievement in AI is so far? Then it gets into Apache Flink and real-time data processing and demonstrating SQL's application in creating simple streaming applications. It also talks about the constraints of using Python for streaming at high volumes and emphasizes the importance of selecting an appropriate language for such tasks.
Now I'm off to think about how I answer what makes me
human...
|
|
|
|
|
|
|
|
|
Join Angela and share your MLOps thoughts!
|
|
|
The Role of AI Safety Standards in Modern MLOps
With so many news stories proclaiming the end of the world because of AI, it can be quite easy to roll your eyes and become blasé about it.
But, when some big dude starts looking through the phone directory for an S. Connor, it'll be too late for "I told you so".
That's where this blog, looking at AI safety and trust, will help. It discusses the importance of embedding safety from the onset of AI development, adhering to international safety standards, and the pivotal role of third-party audits. By focusing on systematic risk management and robust AI governance, you can develop your safety net, rather than Skynet. With thanks to Ritee Rouf from LatticeFlow for their contribution.
|
|
|
|
|
Add your profile to our jobs board here
|
|
|
|
|
|
|
|
|
|
|
|
|
