Murali Sangubhatla

Let’s face it, LLMs aren’t without limitations. In this discussion, Varun Grover shares the benefits of a RAG framework and provides an overview of how it works as well as practical use cases. Watch now: https://vrt.as/43MNQQx #AI #GenAI #LLM #RAG

2

New blog post: The little things that got me into stream processing (https://lnkd.in/gU7Ba5Ym)

60

Honeycomb's storage engine, Retriever, is the key differentiator behind so much of our product. I sometimes get questions from other engineering leaders about why we didn't use x or y off-the-shelf technology or product instead. There are a number of reasons I could share, but one of the most important ones is that it's usually hard to evolve someone else's offering (even if OSS) as the constraints of your business change over time. This post from Hazel Edmands is such a fantastic deep dive into our most recent round of doing just that. https://lnkd.in/g9mvvHWW

121

4 Comments

Just finished recording a great podcast with Ankit Jain on his #HangerDX podcast. We chatted a lot about engineering metrics, their value, and how you should/shouldn't use them to drive change across your organization. It should be released soon, so make sure to check it out! #softwareengineeringinsights #engineeringmetrics #developerexperience

25

Check out the latest in our "Debuggability" blog series! This blog is authored by Kevin Downey Jason Johl and me and talks about how we built a Debuggability Paved Road at Intuit using Ephemeral Containers and Argo Workflows. We'd love to hear your thoughts on the topic. Also, what tools do you build as-a-platform to help debug service, compute and network issues in Kubernetes at scale? #intuit #kubernetes #debug

22

I invite you to join me at the Seattle Postgres 2024 Conference! I’m excited to be one of the speakers, presenting two sessions: 1. Converting Your Schema from Oracle, MSSQL, and MySQL to PostgreSQL in Minutes with DataWharf 2. Tips and Tricks for Developers Moving to PostgreSQL I hope to see you there! #PostgreSQL #PostgresConf #Database #OpenSource #SeattleTech

19

6 Comments

Infusing resiliency into our applications typically introduces complexity. AWS Step Functions (SF) have been my go-to for managing more complex workflows, shifting some of that complexity to a managed service and away from our code. However, adopting SF comes with challenges and may not be a viable option depending on your context. The concept of persistent execution with Temporal is an interesting alternative to consider. You can check more here:

20

Nice example of how you can easily auto scale your containerised #eda applications with #ibmmq and #keda. Taking advantage of MQ's ability to support any number of parallel consumers dynamically

21

Object storage direct-write Kafka topics are coming to Confluent. In data system design, cost is a kind of unstoppable force that rewrites architectures every decade or so, and that is happening right now for high-volume streams. https://lnkd.in/dHRFnrMB

39

1 Comment

"#apachekafka isn't a message queue." Well, that's true for now, but not for long! Currently, Kafka is a distributed log, and it does not have Queue semantics. Log semantics include: - Records are produced to a log. - Records can be consumed, and (within a single partition) they are consumed in the order that they are produced. - Records may be read multiple times by multiple consumer groups. - Records are not "removed" from the log immediately after being read. However, in a message queue, things are different: - Messages are produced to a queue. - Consumers can read and acknowledge messages individually - Messages may be delivered out of order. - Once a message is acknowledged, it is removed from the queue. - Messages may be "retried" if not successfully acknowledged. Kafka is one of the fastest-moving open-source projects. It has an ongoing KIP (Kafka Improvement Proposal) to add Queue semantics on top of its rock-solid log abstraction! Check out KIP-932: https://lnkd.in/gb2Qi__X It will introduce the concept of a "Share Group", which is a new type of consumer group that reads records from a topic with queuing semantics. The Share Group will keep track of which records (well, actually "messages" in this context) have been acknowledged and which ones are still in-flight.

183

7 Comments

In the ever-evolving world of Generative AI, LLMs and SLMs are at the forefront of innovation. With so much buzz around these technologies, it was a privilege to host Sridevi Vadapalli Head of Business Analytics at SitusAMC as part of AIgnite Talk on behalf of Cloud ERP ECF Data Science team. The session provided great insights, shedding light on the importance of choosing the right model for the right job. With rapid advancements in AI, staying informed about these distinctions is more important than ever. Sridevi has dived deep into their differences, best use cases, and the latest developments in the space. LLMs are powerful for complex tasks but require more resources and come with certain risks, while SLMs are efficient and better suited for specialized, quick operations. Choosing the right model depends on the task, balancing the strengths and limitations of each. Rafael Sihorsch Rishabh Tripathi Suman GhoshShuchita Gaur Prakash Kumar Ravikanth MalipedduChristoph Mecke Deep Srivastava Praveen Jha Caleb Dan John Rohan B Sahu. #datascience #ai #ML #saplabsindia #machinelearning #STE #genai #LLM #Innovation #thoughtleadership

117

2 Comments

Our innate ability to fill a blank sheet of paper with our own thoughts makes us more competitive than ever in the generative AI era. Writing is a cultivated skill. It "forces you to slow down, focus your attention, and think deeply." It hedges against incomplete thinking that so easily occurs when we outsource our work to AI as a default. Take a minute to fill half a page with your own stream-of-consciousness. Or, dictate your thoughts to a transcriber app on your phone while you pace (my favorite). It's amazing what ideas you discover are half-baked; what's important/unimportant that seemed so clear only moments earlier in your head. https://lnkd.in/eADNmD78

5

Are you struggling to embrace agility in your organization? ⛓️ Certain behaviors can hinder your team's ability to adapt and innovate. In my latest blog post, I explore common behaviors that can hold your team back and offer practical strategies for breaking free from the chains of traditionalism. Discover how to foster a more agile mindset and drive success. https://lnkd.in/gZSw2CHf #agility #innovation #leadership #changemanagement #teamwork

3

Announcing our second conversation, Raza Habib with Shreya Shankar! The LLM mini-conference to equip you to build: "So you think can prompt" June 22 in SF. Sign up here: https://lu.ma/iulmro47 Co-hosted with Hamel H. and Slater Stich at Bain Capital Ventures. This one is going to be two experts discussing what's real and what's nonsense in evaluation. This conversation is going to be gold.

25

1 Comment

This week, the Pacific Northwest weather gave us an unplanned dose of "Yutori." Power outages, no internet, and suddenly, the hustle and bustle of life came to a screeching halt. In essence, Yutori suggests creating space in one's schedule and mindset for a slowdown, a blankness, allowing oneself to savor the sights and sounds around you. When you don’t rush; there is no concern for time. The flickering of a candle, a child’s laughter, my mothers’ voice on the phone. These are things that, in my usual hectic schedule, I barely have time to relish. But this week, as we sat in the darkness, we played games, read books, and reconnected with friends. Living with Yutori means embracing spaciousness. It’s about balancing work and personal well-being, without the constant urge to do more. Yutori can manifest in various ways, rooted in setting boundaries and sticking to them. It’s a fresh take on the often-overused concept of “work-life balance.” Let’s be honest, the phrase "work-life balance" has become so cliché that we almost roll our eyes whenever we hear it. But doesn't it feel good to use a term that conveys the same idea, but with a bit more flair and substance? Creating those boundaries isn’t easy, especially in a world where "do more with less" seems to be the motto. Even with the flexibility of working from home, many of us still find ourselves stretched thin. Efficiency and productivity might be up, but so is exhaustion. Despite having more control over our schedules than ever before, we often feel less in control and more overwhelmed. What I’ve gradually come to realize is that what’s missing from our jam-packed calendars is a sense of play, of ease, and spaciousness. There is unexpected beauty in slowing down, and joy in savoring simplicity. As a wise colleague once said - No one else is going to give you that lunch break , M- its you! :-)

19

I am excited to share the latest blog, co-authored by myself, Junhwi Kim and Vyas Sekar, detailing how we managed to successfully localized and improved LLM by combining Fine-Tuning and RAG. https://lnkd.in/gKDkgGPq

70

3 Comments

In this panel session, explore the trends and prediction of Open Source technology and how OCI supports Open Source workloads with differentiating performance, security, cost benefits through innovation. https://lnkd.in/g-UekgEC

2

1 Comment