How 'real time' is real time?

MLOps vs ML Orchestration // Ketan Umare // MLOps Podcast #182

As a much smarter man than me said, “Time is relative”.

And, yeah, he may have been talking about the big grand scheme of things, but we see it in our lives all the time.

Sitting in the dentist’s chair for 5 minutes can feel like hours.

Your week-long vacation feels like it’s over in a day.

Trying to implement real-time in your build. I mean, what even is real-time?

5 minutes? 1 minute? 1 second?

Well, in this episode Ketan Umare from Union AI talks about how it’s all relative to the model build.

He also shares insights on the importance of reproducibility in MLOps, impact of data drift, his perspective on the evolving role of MLOps engineers, and how Union helps with all of these.

Just make sure you give yourself enough time to enjoy it.

Video || Spotify || Apple

Data Platforms in MLOps: Translating Business Goals into Product Decisions // Richa Sachdev // MLOps Podcast #184

You clean and process your data, you go through cycles of train, evaluate finetune.

You deploy.

It runs like a dream.

It’s the perfect model! Time to scale up!

Except, the end-user experience sucks and it doesn’t match the business goals.

It can be easy to lose sight of these when building your model.

Richa Sachdev gives a good insight in to this. She understands knowing the tech is valuable, but more valuable for her now in her role leading Data Reporting and Process Automation at JPMorgan Chase, is knowing the business side of things.

She has to understand the business goals and metrics that they're hitting.

Then she can translate it into the product design decisions and the technology decisions to optimize the end-user experience.

And part of that is heling me out with those free flight up-grades…

Video || Spotify || Apple

Job of the week

6 Software Engineers // Fermata Energy (remote but US based)

Fermata Energy leads V2G innovation, harnessing parked EV batteries to enhance grid efficiency and reduce customer costs.

We seek a backend developer for our IoT services for V2G applications on Amazon AWS. Software projects range from highly scalable streaming services to hardware software integration.

Need to have skills:

Two plus years experience contributing to real-world software development
Understanding of software development best practices
Experience developing applications on SQL and/or NoSQL databases and with distributed services and/or messaging systems
Proficiency with at least one cloud platform such as AWS, GCP, or Azure
Understanding of API design and best practices

Craft, Scale, Win: Your ML Platform Journey! // Andrei Vishniakov // Meetup IRL #50 Berlin

You want to watch a horror film, they want comedy.

You want pizza, they want Chinese.

What do you do? Compromise and watch Shaun of the Dead with a pizza topped with chow mein.

Nothing quite as drastic in this talk by Andrei Vishniakov, a Machine Learning Engineer at ZenML, but he does look at the need for compromise.

He highlights the need for better collaboration and compromise between data science and engineering teams and explores the alternatives of using all-in-one platforms and best-of-breed tools, discussing the trade-offs and benefits of each approach.

Watch it here

Dive Deep with Notion, Snowflake & Qwak on Nov 8th! 🚀

🗓️When? November 8th, 11:30 AM EST
🎙️Who? Notion, Snowflake & Qwak

Why Should You Be There?

Discover the mechanics of Notion's Search & Ranking ML Model. This isn't just tech chatter—it's the key to the marvels of a leading productivity tool.

🤝See Snowflake’s Data Cloud and Qwak's ML platform in action, showcasing a unique collaborative integration.

Highlights:

Unravel the genius behind Notion's ML in enhancing search and ranking.
Delve into how ML refines user experiences, ensuring efficiency & ease.
Witness how Notion utilizes Snowflake's might to serve a massive audience flawlessly.
Glimpse into the exciting future of ML in reshaping productivity software.

Interested in AI and ML, a fan of Notion or Snowflake, or simply eager to learn? Don't miss out! Join this session to directly engage with and learn from experts at Notion, Snowflake, and Qwak.

🔗Reserve Your Seat here!

Conversational Memory in LangChain with Milvus

Walking in a room and forgetting why.

Looking at your watch and forgetting the time straight away.

Birthdays and anniversaries.

Starting a list of things and forgetting why…

Oh yeah, memory, of course!

It can be tricky for humans, never mind LLMs, but this blog by Yujian Tang, a Developer Advocate at Zilliz, should at least help with Conversational Memory with LangChain.

In it, he covers:

Conversational Memory with LangChain
Setting Up Conversation Context
Prompting the Conversational Memory with LangChain
LangChain Conversational Memory Summary

Click the link below, just don’t forget to read it!

Read it here!

ML Proverbs by Leonard Aukea

The chow mein-topped pizza mentioned above probably wouldn’t come with a fortune cookie, but worry not.

You can get your ML pearls of wisdom from this insightful blog.

Curated by Leonard Aukea, he offers up some often-quoted guidance and ideas as ML Proverbs, including:

Measure twice, optimize once.
Clear is better than clever.
Untested assumptions become unseen errors.

He’s also keen to continue to develop it, so if you have any other insights, let him know.

Be sure to read it here and enjoy with a take-away of your choice!