על אודות
פעילות
3K עוקבים
-
Yedidya Schwartz שיתף את זהOn November 5 I’m speaking at Conf42 Platform Engineering. Like many of us, I’ve been working on improving our AI SDLC. Some things worked, some failed badly, and I tried a lot of tools and strategies along the way. I collected a lot of lessons, and I’m excited to share them. I looked into what really matters in an AI SDLC (spoiler: writing the code is not the main part), and how big companies like Stripe, Spotify and Uber are doing it. You won’t leave with a perfect formula for a software factory. But you will leave with new ideas and some inspiration. It’s online and free, so join me. Link in the comments. Conf42 Mark Pawlikowski
-
Yedidya Schwartz שיתף את זהIt was a privilege to share Quicklizard's agentic vision at the Executive Gen AI Circle at Cramim Hotel. A great event with inspiring conversations. Thank you for having me
-
Yedidya Schwartz שיתף את זה𝗪𝗲 𝗷𝘂𝘀𝘁 𝘀𝗵𝗶𝗽𝗽𝗲𝗱 𝗼𝘂𝗿 𝗠𝗖𝗣, 𝗮𝗳𝘁𝗲𝗿 𝗮 𝗳𝗲𝘄 𝘄𝗲𝗲𝗸𝘀 𝗼𝗳 𝗶𝗻𝘁𝗲𝗻𝘀𝗶𝘃𝗲 𝘄𝗼𝗿𝗸. Strategies, competitor prices, price history, performance. All of it is now available through natural language, instead of searching for it dashboard by dashboard. Happy to be leading this and bringing cutting edge technology into the retail industry. More details in my latest article, link in the comments.
-
Yedidya Schwartz שיתף את זהI spent weeks making AI agents build software on their own, and some of that time was spent watching it break. I see many posts about the shiny agent demos, and not so many about how not to do it... So here is the other side, I'm working on turning our whole software development lifecycle into a durable, self-driving pipeline on Temporal on the cloud (no need to have your laptop opened), and reality isn't so simple. It started as skills we ran by hand, step by step, and I decided to take those manual steps and turn them into a deterministic workflow orchestrated by Temporal. Every pain point is in here: the disk that lies the second you scale, the metric that says zero while the agent is clearly working, the observability that went dark for few days, the agents that swear the build is green when nothing compiles. I recommend reading it before you start, not after it burns you. That project is still work in progress, and probably will stay that way as this domain keeps evolving every day, but I though that sharing some insights now would be helpful to some of you. Link to the article in the comments
-
Yedidya Schwartz שיתף את זהJust wrapped up an inspiring week at Developer World - DWX 2026 in Mannheim, Germany - the conference for AI, Cloud, Web, and more. Proud to see Yonatan Kra, our Principal Architect at Quicklizard, give a great talk on how the tricks the gaming industry uses for speed and smooth user experience can be applied to everyday apps. A few other talks I attended that were very insightful, Patrick Koss from STACKIT - on many eye-opening distributed system architecture challenges, telling stories from real experience with complex systems. Daniel Ostrovsky from Payoneer - a refreshing look at how MCP tools can easily leak your data or run malicious code, and why blind trust in AI tooling is super risky. Alon Fliess from ZioNet - on letting an AI agent build an entire cross-platform game with zero human-written code, and why the future is spec-driven development, where you architect and the agent builds. Marin Niehues from Stage Starter - a real-world story of implementing AI at a global manufacturer, where the biggest obstacle wasn't the tech or the budget, but the people.... office politics, skeptical managers, and deeply rooted habits. A big thank you to the organizers and speakers for such an enriching week. #DWX2026 #AI #conference
-
Yedidya Schwartz שיתף את זהHad a great time joining Dr. Fabian Uhrich on the first episode of the PriceCheck podcast. We covered how AI is changing retail from multiple angles - from customer decision-making to pricing infrastructure. One idea I especially enjoyed discussing: pricing can no longer stay periodic or average-based. Every SKU needs its own context, signals, and decisions - continuously. Hear about this and more in the episode, link in the comments #PriceCheck #RetailAI #PricingStrategy Quicklizard
-
Yedidya Schwartz שיתף את זהIn my latest article I explain why Agentic Commerce Optimization starts before the shopper ever lands on your site:Yedidya Schwartz שיתף את זה𝗜𝗻 𝟮𝟬𝟮𝟲, 𝘆𝗼𝘂𝗿 𝘄𝗲𝗯𝘀𝗶𝘁𝗲 𝗶𝘀 𝗻𝗼𝘁 𝘆𝗼𝘂𝗿 𝗳𝗿𝗼𝗻𝘁 𝗱𝗼𝗼𝗿. 𝗬𝗼𝘂𝗿 𝗱𝗮𝘁𝗮 𝗶𝘀. 🤖✨ In the era of agentic commerce, AI agents are increasingly making purchasing decisions before shoppers ever reach your site. That means pricing accuracy, feed freshness, and real time data consistency are no longer operational details. They are competitive advantages. A delayed price update is no longer just a technical issue. It is a lost sale. In his latest blog, Yedidya Schwartz explores why pricing data fidelity is becoming the foundation of Agentic Commerce Optimization (ACO), and why retailers need to rethink pricing as a live trust signal rather than a static number. Worth the read for anyone thinking about retail, pricing infrastructure, and the future of commerce. 👉 Check it out here: https://lnkd.in/dN-waMnf #AgenticCommerce #AICommerce #DynamicPricing #RetailTech #Ecommerce
-
Yedidya Schwartz שיתף את זהJust got the email that I'm continuing into my third year as an AWS Community Builder. Looking forward to contributing where I can and learning along the way Amazon Web Services (AWS) #awscommunitybuilders AWS User Group Israel
-
Yedidya Schwartz שיתף את זהx402 enables true pay-per-consumption at the protocol level, and I think this might shift products pricing strategy for everyone. Read more in my latest articleYedidya Schwartz שיתף את זה𝗔𝗜 𝗮𝗴𝗲𝗻𝘁𝘀 𝗰𝗮𝗻 𝗮𝗹𝗿𝗲𝗮𝗱𝘆 𝗳𝗶𝗻𝗱 𝗽𝗿𝗼𝗱𝘂𝗰𝘁𝘀. 🤖 ⚙️ 𝗡𝗼𝘄 𝘁𝗵𝗲𝘆 𝗰𝗮𝗻 𝗽𝗮𝘆 𝗳𝗼𝗿 𝘁𝗵𝗲𝗺. For the past two years, agentic commerce has been focused on discovery. But discovery without payment is just a recommendation engine. That’s what makes x402 such a shift. In his latest article, Yedidya Schwartz, CTO at Quicklizard, breaks down how x402 introduces a native payment layer for AI agents, allowing them to complete transactions instantly, without accounts, checkouts, or human approval. And that changes more than payments. It changes how competition works. When agents transact on your behalf, there’s no brand story or UX to influence the decision. Just price, performance, and reliability, evaluated in milliseconds. Which means: • Your pricing becomes a real-time commitment • Your data accuracy becomes a transaction requirement • Your infrastructure becomes your storefront And for the first time, micropayments work at scale. The takeaway is simple: the “buy button�� is disappearing. The systems behind it are the real edge. 👉 Read the full article here: https://lnkd.in/dhw_w8ME #AgenticCommerce #AIinCommerce #Fintech #RetailTech #Micropayments #Ecommerce #DigitalPayments
-
Yedidya Schwartz אהבתי את זהYedidya Schwartz אהבתי את זהWe're #hiring a new Technical Support Engineer (Tier 2) -Team lead in Petah Tikva, Center District. Apply today or share this post with your network.
-
Yedidya Schwartz הגיבו על זהYedidya Schwartz הגיבו על זהWe’re thrilled to welcome Amit Zarmon and Maish Saidel-Keesing to the Solutions Architecture team at Automat-it! Amit joins us as a Solutions Architect, and Maish as Senior Solutions Architect Team Lead, both in Israel. We’re excited to see what great things they’ll accomplish with us. Welcome to the team! Let’s make amazing things happen together. #Welcome #NewTeamMember #Automatit
-
Yedidya Schwartz אהבתי את זהYedidya Schwartz אהבתי את זהI'm excited to join Automat-it as Solutions Architect. Looking forward to working with an incredible team, helping startups grow with AI, DevOps, and Data on AWS, and contributing to the next stage of the company's growth. Looking forward to the journey ahead! 🔥 💃 Nir Shney-Dor Maish Saidel-Keesing Automat-it
-
Yedidya Schwartz אהבתי את זהYedidya Schwartz אהבתי את זהנגמרים לכם הטוקנים בקלוד קוד מהר מדי? יכול להיות שמצאנו את אחד האשמים. בשבוע האחרון הרשת (ומתי בעצם לא?) התמלאה בתלונות של משתמשי Claude Code שטענו שמכסת השימוש שלהם נגמרת הרבה יותר מהר מהרגיל. חלקם אפילו דיווחו על צריכה גבוהה פי 5 ממה שהכירו. בוריס צ’רני, היוצר של קלוד קוד, החליט לבדוק את העניין וביקש ממשתמשים לשלוח לו Session IDs כדי שהצוות יוכל לחקור. אחרי שעברו על הסשנים, הם מצאו כמה באגים קטנים שגרמו לצריכת טוקנים מיותרת (ותיקונים כבר בדרך), אבל לא מצאו רגרסיה אחת שמסבירה את כל התלונות. מה הם כן גילו? אחד הגורמים המרכזיים הוא השימוש בפיצ’רים כבדים כמו Agent Teams, Workflows ורמות Effort גבוהות. ופה מסתתר פרט קטן שהרבה משתמשים כנראה לא מודעים אליו: כברירת מחדל, Subagents יורשים את רמת ה-Effort של הסשן הראשי. כלומר, אם אתם עובדים עם xhigh effort וקלוד מחליט להפעיל כמה Subagents, גם הם ירוצו ב-xhigh. אפילו אם כל מה שהם צריכים לעשות זה לחפש קובץ או לבדוק משהו פשוט בקוד. וכשיש כמה כאלה במקביל, הטוקנים יכולים להישרף בקצב מרשים במיוחד. 🔥 אז משתמש אחד שאל את בוריס שאלה מתבקשת: למה שלא נוכל להגדיר ל-Subagents רמת Effort נמוכה יותר כברירת מחדל? ויום למחרת בוריס כבר הכריז על משתנה סביבה חדש: CLAUDE_CODE_SUBAGENT_EFFORT עכשיו אפשר להפעיל את קלוד קוד כך: CLAUDE_CODE_SUBAGENT_EFFORT=low claude וזהו. ה-Subagents ישתמשו כברירת מחדל ב-Effort נמוך, בלי לשנות את רמת החשיבה של הסשן הראשי. אפשר גם להגדיר את המשתנה באופן קבוע ב-settings.json, או לקבוע רמת Effort ספציפית לכל Custom Agent. כמובן, לא כל משימה מתאימה ל-low effort. לפעמים אנחנו דווקא רוצים שה-Subagent ישקיע יותר מחשבה, אבל לפחות עכשיו יש לנו שליטה טובה יותר על זה. ועוד טיפ קטן: הפקודה /usage יכולה לעזור לכם להבין לאן הטוקנים שלכם הולכים. אז בעיניי, לא מדובר פה רק על משתנה הסביבה החדש, אלא העובדה שכלי שהופך ליותר ויותר אוטונומי יכול גם לצרוך הרבה יותר משאבים מאחורי הקלעים, בלי שהמשתמש באמת מבין למה. וזאת כנראה בעיית UX לא פחות מבעיית עלויות. מעניין כמה טוקנים הייתי יכול לחסוך אם ההגדרה הזאת הייתה קיימת קודם. 😅 יש לי עוד כמה טיפים ודעות על סאב אייג׳נטס, אולי אפרסם עוד השבוע.
-
Yedidya Schwartz אהבתי את זהYedidya Schwartz אהבתי את זהKarpathy’s rule is simple: prompts are now programs that program the LLM. And they are written in English. That sounds casual. It is not. If English is the programming language, then the prompt is production code. Not a sticky note. Not a chat you forgot to save. When teams ignore that, the agent looks fine in a demo and drifts in production. Tone changes. Edge cases slip. Nobody can say which prompt shipped last Tuesday. What I keep coming back to when building voice, WhatsApp, and CRM agents: → Version every system prompt the way you version code → Write evals for the behaviours you care about before you ship → Review prompt changes like pull requests, not Slack edits → Monitor live outputs for drift, not just uptime → Keep a rollback path when a “small tweak” breaks the agent Software 3.0 does not remove engineering. It moves more of it into English. Treat prompts like code, or the system will remind you they were. How are you versioning and testing prompts in production today? Credit: Andrej Karpathy, YC AI Startup School
-
Yedidya Schwartz אהבתי את זהYedidya Schwartz אהבתי את זהקורס פרונטלי נותן משהו שכבר קשה למצוא כיום: מסגרת שבה אנשים מתגבשים ויוצרים קשרים זה עם זה. מדי פעם אנחנו נפגשים לאירוע פנימי של קהילת הקורסים (הבסיסי והמתקדם) להמשך למידה (כי ללמוד זה סטייט אוף מיינד), וביום חמישי האחרון נפגשנו סתם כדי לשתות. האמת היא שהסיבה הפורמלית היתה שגם אריאל Ariel Shulman וגם אני התחלנו עבודה חדשה. אבל תכלס נראה לי שגם בלי סיבה פורמלית היה ממש כיף לשבת ולקשקש קצת. אה, כן, לא התאפקתי והכנתי טאבו בקאנד (פעם הבאה - אליאס)! איזה כיף שיש קהילה של אנשים שפשוט נהנים להתפתח מקצועית! 👏 החל מיום ראשון - נפתח הרישום למחזור 4. יאיי! 🥳 לאתר הקורס ולתוכנית הלימודים (שהתארכה קצת!): https://lnkd.in/edKHQdse
ניסיון חינוך
רישיונות והסמכות
ניסיון בהתנדבות
-
Emotional First Aid
ERAN Emotional First Aid by Telephone & Internet
-להציג 4 שנים 7 חודשים
שירותי רווחה
-
Patrolman
Israeli Police
-להציג 7 שנים 11 חודשים
שירותי רווחה
Maintaining public order and public security, law enforcement, and the fight against crime. In his capacity, the patrol addresses police incidents such as handling domestic violence incidents, serious criminal incidents, and more.
-
Board Member
Eshnav
-להציג 10 שנים 9 חודשים
מדע וטכנולוגיה
Board member at Eshnav - People for smart use of the internet
https://www.eshnav.org.il/
-
Cleanliness Trustee
Ministry of Environmental Protection
-להציג 6 שנים 7 חודשים
סביבה
Cleaning trustees may, by virtue of their training and authority, assist law enforcement through a report to the Ministry of the Environment on offenses like dumping debris from a vehicle etc.
http://www.sviva.gov.il/English/env_topics/cleanliness-trustees/Pages/Littering-Fines.aspx -
Editor
Wikipedia, the Free Encyclopedia
-להציג 6 שנים 11 חודשים
השכלה
Language editor at the Hebrew Wikipedia
פרסומים
-
Using a Buffer Microservice and Amazon SQS to Reduce the Load on a DB
WalkMe Engineering blog
ראו פרסוםIn this article, I describe how to prevent and test system overload — without increasing any resources.
-
Horizontal Scaling of a Stateful Server with redis pub/sub
WalkMe Engineering blog
ראו פרסוםIn this article, I wrote how we solved the issue of horizontal scaling in a stateful monolith to give our good old editor server some more time before the sun sets on it.
-
Meetup lecturer - Microservices architecture case study
Walkme Tech Talks
ראו פרסוםhttps://www.meetup.com/WalkMe-Tech-Talks/events/267879116/
Microservices have been a buzz word for a long time now. In essence, it is an architecture that helps manage complex systems by splitting them into small pieces.
Most systems begin as a monolith and the need to break it down comes much later in development.
While this might be a big and scary project, we have recently migrated big parts of our monolith to microservices and lived to tell the tale.
In this talk I…https://www.meetup.com/WalkMe-Tech-Talks/events/267879116/
Microservices have been a buzz word for a long time now. In essence, it is an architecture that helps manage complex systems by splitting them into small pieces.
Most systems begin as a monolith and the need to break it down comes much later in development.
While this might be a big and scary project, we have recently migrated big parts of our monolith to microservices and lived to tell the tale.
In this talk I shared the process that helped us split a big monolith into a microservice gradually and with zero downtime or failures.
קורסים
-
BI + ITIL
2868
-
Creative and advertising psychology
772
-
Introduction to internet networks
5850
-
Officers course
-
-
Vital Signs Advanced - Microsoft windows server
-
-
Vital Signs: Performance Monitoring Windows Server
-
שפות
-
Spanish
מיומנות שפת אם או דו-לשונית
-
English
מיומנות מקצועית בעבודה
-
Hebrew
מיומנות שפת אם או דו-לשונית
ארגונים
-
IDF
-
-
המלצות התקבל
3 אנשים המליצו Yedidya
הצטרפו עכשיו כדי נוף