Adi Glasman

🚀🎙 Episode 25 of #PeymanFakharianPodcast and this time with Kayvan Mohit Mafi, CEO of Diginext.ir, the innovation center of #digikala. Kayvan truely is one of the most active supporters of startup innovation in Iran and in this episode he shares his insightful views on B2B tech and AI startups. 🔍 Topics We cover: - Foundation of Diginext - Story of Diginext and its investment portfolio - Investment on B2B Startups - AI Startups - Inviting Mentors to help startup founders 🎧 Listen to the episode on your favorite podcast app. Just search for "Peyman Fakharian" or "پیمان فخاریان" 🎧 Or listen on Castbox https://lnkd.in/eRQ6AmzA 🎧 Or Subscribe to our Youtube Channel https://lnkd.in/ePx6cfS7 📌 For more updates and exclusive content, follow my Telegram Channel: https://lnkd.in/dQJVz2Z9 hashtag #Peymanfakharian #پیمان_فخاریان #digikala #diginext #دیجیکالا #دیجی‌کالا #PeymanFakharianPodcast

129

6 Comments

Confluent's Field CTO, Kai Waehner, shares insights on the future of streaming analytics and data management, highlighting the importance of reliable data flow, collaboration, and security. Discover how Confluent's cutting-edge platform revolutionizes streaming analytics. Read the full interview: https://lnkd.in/gSMVCmvy #StreamingAnalytics #DataManagement #FutureOfData #BusinessIntelligence

80

Just because your service and or data is in the cloud doesn't mean that its going to be available. You need to understand what is meant by shared responsibility.

1

What are your early career advice? I had a good time chatting with Sam on this. Check out the video below. You can also post your advice in my comments to share. =) https://lnkd.in/g7ZfE8y3

31

2 Comments

My #opensource history has a lot connection to where the water spins in the other direction, and #golang. Chris Mazur dared to host my first talk on #webassembly a couple years ago, in Wellington. This was on #wazero.. pure go runtime (against the odds in a rust ecosystem). This year, Katie Fry dares to host my first talk on #genai in go (against the odds in a python ecosystem), in Sydney. To write a first talk for me is a mammoth task. I get really nervous and do tons of research, ask my friends, look at a million demos. I have 300 browser tabs, tons of example code locally, several rambling text editor windows. Co-located with an Elastic meetup, my presentation at Katie's is all about Go, and I feel good about it. I had a quite decent plan together, even if later versions go farther. Yesterday, I had some initial scratches towards a real outline. This morning, Philippe Charrière pointed me to his 3 month old baby, https://lnkd.in/eTCRpNPr. This is one of his finest works: it is another improvement on his developer empathetic approach to #opensource. His examples are simple, tiered and top-notch. As GenAI is such a vast and drifty space, this is even more important for noobs. Of course, I was unsurprised this integrates with #webassembly functions via Extism. So, when you ask a question in a way the LLM understands a tool is relevant, those functions can be in any language, not just #tinygo, but anything that works in wasm, like rust. I chuckled stumbling on this, unsurprised as I worked with Phillipe earlier with his capsule FaaS project. TL;DR; I basically threw out my examples and will start over and use parakeet. This doesn't imply you will end with parakeet, but frankly there are far worse ways to start in Go. Who knows where you will end, but starting is the most important part. I'm hoping folks in Sydney enjoy this as much as I do, and also hope this turns into another successful way to relate a new tech, yet again starting where the water spins in the other direction. https://lnkd.in/ebZfF9Va

23

A very enlightening perspective on whether the "juice is worth the squeeze" for a B2B company to start a product-led growth effort and what their journey would be like (from minute 45 - 55): https://lnkd.in/eTWY6CRW). I love how Shaun Clowes also discusses the early days (2012!) of PLG.

1

I have been leveraging GCP for TrussHealth's AI and data storage, along with addressing customer data related issues in previous roles. Google's transparency in addressing and resolving recent data loss incidents instills confidence in their commitment to customer support and functional operational processes. Although data loss is a significant concern for enterprise storage, it is commendable to see cloud providers like Google openly acknowledge such challenges and demonstrate dedication to continuous improvement. Learn more about how Google Cloud responded to accidental data deletions: [Link to the article](https://lnkd.in/gTiRkg6X)

19

Very much needed analysis, good to see top models across the measures.

10

My beginner's tutorial has been shared by LangChain! 👍 Personally, I am new to large language models (LLMs) as well, and it's been such a fascinating journey exploring this field as LangChain makes it so easy to utilize these models and start developing. If you're interested in any of Python, LangChain, AI, LLMs, or DragonflyDB, check it out! Your support and feedback would mean a lot to me. 📚✨ #AI, #GenAI, #LangChain, #LLM, #OpenAI, #Python

14

3 Comments

Our company recently received an additional $75,000 in AWS credits, which will extend our runway even further. This incredible support is largely due to our brilliant CEO and co-founder, she is one of the best managers I’ve had the privilege of working with here in Sydney - the company leader you want to invest in. Of course, we also owe thanks to the AWS team for their exceptional service and their confidence in our small team and unique AI value proposition. As for me, I’m just the CTO—the one clicking the buttons and hoping they’re in the right order to get results! :-) This post isn’t just about celebrating our progress but also about expanding our network of investors and connecting with other dedicated co-founders here in Sydney. Please message me or connect with me on LinkedIn if you’d like to connect! I also want to share some insights on two popular topics within our community: (1) finding a co-founder, and (2) building solid tech architecture. 𝗙𝗶𝗻𝗱𝗶𝗻𝗴 𝗮 𝗖𝗼-𝗳𝗼𝘂𝗻𝗱𝗲𝗿: In my journey, I interacted with many people before finding the right co-founder to work with. Some green flags to look for include strong execution skills, business acumen, a focus on results, great communication, consistency, and kindness. Red flags, on the other hand, include sporadic or slow communication, low integrity, lack of market understanding, and unreasonably high demands / low offer. Ultimately, your co-founder should share core values and a mission aligned with your own. 𝗧𝗲𝗰𝗵 𝗔𝗿𝗰𝗵𝗶𝘁𝗲𝗰𝘁𝘂𝗿𝗲: AWS credits are invaluable, but they can’t compensate for poor architectural choices or bad tech decisions. Missteps here will quickly consume credits, leaving you with technical debt that may set you back further than where you started. Here are some architecture tips: 𝟭. 𝗞𝗜𝗦𝗦 (𝗞𝗲𝗲𝗽 𝗜𝘁 𝗦𝗶𝗺𝗽𝗹𝗲, 𝗦𝘁𝘂𝗽𝗶𝗱): Aim for functionality over complexity—don’t try to design the next SpaceX shuttle! 𝟮. 𝗦𝘁𝗮𝘆 𝗢𝗿𝗴𝗮𝗻𝗶𝘀𝗲𝗱: Clear, maintainable code and a logical project structure will save time and reduce errors over the long term. 𝟯. 𝗦𝗼𝗹𝘃𝗲 𝗣𝗿𝗲𝘀𝗲𝗻𝘁 𝗣𝗿𝗼𝗯𝗹𝗲𝗺𝘀 𝗙𝗶𝗿𝘀𝘁: Avoid over-engineering for hypothetical future needs; focus on solving the immediate issues at hand. 𝟰. 𝗨𝗻𝗱𝗲𝗿𝘀𝘁𝗮𝗻𝗱 𝘁𝗵𝗲 𝗥𝗲𝗮𝗹 𝗣𝗿𝗼𝗯𝗹𝗲𝗺: Ensure you’re addressing the core problem instead of copying existing solutions that may not be relevant. 𝟱. 𝗣𝗿𝗶𝗼𝗿𝗶𝘁𝗶𝘀𝗲 𝗕𝘂𝘀𝗶𝗻𝗲𝘀𝘀 𝗮𝗻𝗱 𝗣𝗲𝗼𝗽𝗹𝗲 𝗢𝘃𝗲𝗿 𝗧𝗲𝗰𝗵𝗻𝗼𝗹𝗼𝗴𝘆: Technology serves the business and its users, not the other way around—keep your decisions grounded in this. 𝟲. 𝗧𝗵𝗲𝗿𝗲’𝘀 𝗡𝗼 "𝗕𝗲𝘀𝘁 𝗧𝗼𝗼𝗹"—𝗝𝘂𝘀𝘁 𝗕𝗲𝘁𝘁𝗲𝗿 𝗨𝘀𝗲 𝗖𝗮𝘀𝗲𝘀: Tools are context-dependent, and skill level matters more than the tool itself. The best developers know how to adapt the right tool to the right context. Please if you connect with me make sure you are inside Australia, ideally in NSW. #cto #sydney #startup #technology #ai #machinelearning

35

7 Comments

Hear from Peter Hanssens about an interesting feature of Rust for data engineering. Spoiler alert: portability and distribution matter. As well as being the lead organiser of the ServerlessDays ANZ conference, he runs Cloud Shuttle. I was able to grab a few minutes of his time during the conference today to check in on how the course Accelerant provided earlier this year has been able to solve some problems for their customers. #rustlang #dataengineering

45

2 Comments

Text-to-SQL with GenAI isn’t the magic bullet some vendors make it out to be. It’s tempting to believe the hype, especially after watching demos that answer simple questions like "What is the population of Australia?" on well-curated datasets - but the reality is far more complex. I was glad to hear several speakers at the #dataengbytes24 conference share similar experiences. One of the standout talks for me was by Juan Sequeda from data.world, who analysed how well Text-to-SQL actually performs. He broke down queries into four categories, ranging from low-context, simple metrics (like sums and averages over a single table) to high-context, complex metrics (involving multiple joins and intricate logic). The numbers spoke for themselves: overall accuracy was only about 16%, and for high-context queries, it dropped to 0%. Some key reasons for incorrect queries include hallucinations - generating non-existent column names, incorrect values applied as filters, and inaccurate joins. Juan’s proposed solution was to incorporate knowledge graphs into the process. By using this structured layer of context (Text-to-KG-to-SQL), accuracy improved by as much as 4x, making even complex, multi-join queries far more reliable. Knowledge graphs provide the critical depth and structure that GenAI needs to deliver meaningful results. It’s also important to be aware of other challenges: ambiguous queries (different interpretations of the same thing), unanswerable questions, default assumptions, and multi-level aggregations (multiple questions in one). These are just a few of the common pitfalls when using GenAI for SQL.

91

13 Comments

The next Sydney Platform Engineering Meetup is locked in for July 4th! The sessions lined up for this meetup are 🔥 Neha Boora from Class Pty Limited is going to share how to operationalize AI and ML for managing cloud resources. Neha will cover topics like enhancing observability with AI, using AI and ML to predict future workloads, and adjusting resource allocation based on real-time demand patterns. Jovana Dunisijevic spent years working on products like Atlassian Jira before joining the Atlassian Platform Engineering team. In this Q&A session, Rad Macakanja from Amazon Web Services (AWS) will ask Jovana a series of questions exploring the differences in product and platform engineering, and what to consider if you're thinking about making a similar move. We're back in the Atlassian Sydney office for this free event! Tickets are limited (we sell out every meetup), so get yours now! The link to RSVP is in the comments. #DevOps #PlatformEngineering #AI

77

8 Comments

When ANZ buyers were asked why they are replacing software, it's very interesting that Security is driving these changes (and syncs with what we see in other regions). Security was the #1 purchase consideration from our 2024 Buyer Behavior Report. Question: Why does your company usually replace software? Select up to 3.

12

Next Melbourne Compose meeting is 6pm Thurs Oct 17. In-person@Kathleen Syme Centre, Carlton https://lnkd.in/gJazYCz3 john walker - Tooling around with Haskell: "So you've learned the Haskell basics and now you want to build something. You'll need to find and install dependencies, resolve versioning issues, type check, navigate, build and run your code, but how? Haskell tooling has come a long way over recent years, but there are still multiple paths to achieving the above. Stack, Stackage, Cabal, cabal-build, hpack, ghci, ghcid, ghcup, HLS and recent newcomers Static-ls and ghciwatch can seem like an alphabet soup of complementary and competing tools to the uninitiated. This discussion will be part history lesson, part experience report and part demo. I'll give an outline of the development timeline for some of these tools, describe how and why I've landed on the tool set I have and give a brief demo. To the Haskell curious, you'll come away with an understanding of the capabilities of these tools and how to use them. For experienced Haskellers, please come and fill us in on how your setup works. For Nix users, please be outraged to the point of volunteering to take a future session." Hope to see you there, Ben

8

4 Comments

Here's a great article describing Weaviate's guidance on a process called "late chunking". RAG can get frustrating when embeddings are stored in a way that separates critical contextual relationships between sentences, paragraphs, or entire documents. Late Chunking helps to address that problem. With more language models providing larger context windows, the prompt augmentation strategy should also evolve. I like what they're doing... https://lnkd.in/gq-4vQBg

9

1 Comment

Ever wondered how AI works? 🤔  Today, on the @DayofAI, CSIRO and Google Australia unveil a new video to explain just that. Learn how AI is being used across Australia – from healthcare to agriculture – and why responsible AI is crucial for a fair and inclusive future. Watch the video and join the conversation! #AI #ResponsibleAI #DayofAI #CSIRO #GoogleAustralia https://lnkd.in/giWhMT5s

19

I attended Kubernetes Community Day at the ICC. I met a lot of familiar faces and had a lot of fun. Some take aways: -Some talks were interesting while others very basic. A bit of a heads up would be appreciated. -ANZ+ is leading the way with some sophisticated uses of Kubernetes while other organisations are just starting. Is interesting to see the adoption continuum. -Dear CBA, why do you need almost 2000 clusters? Isn’t that a bit too much and a lot of overhead to replicate the control plane so many times? Also, there was a slide mentioning Platform Needs vs Client Needs and the speaker said they both have the same priority. I understand that without a client there is no platform. How exactly clients don’t come first? -Argo CI/CD is great, but how about the others Argo X products. How they stand in the market. Why should people should use them? -Food and venue were great. I’ll be back.

5