Traceloop

What if prompt engineering wasn’t the future of enterprise AI? Could automation and test-driven workflows replace trial-and-error with repeatable results? ⚡ Nir Gazit, CEO and Co-Founder of Traceloop is redefining how organisations deploy and scale large language models (LLMs) reliably. With over 15 years in software development and machine learning—including leadership roles as Tech Lead at Google and Chief Architect at Fiverr—Nir now leads Traceloop in building tools for real-time detection, evaluation, and remediation of anomalies in AI systems. He also co-leads the Generative AI Special Interest Group at OpenTelemetry, advancing open standards for AI observability. 💥💥💥 Based in Tel Aviv, Nir is known for bridging cutting-edge AI research with practical enterprise solutions. Join him at the use case session: “Prompt Engineering is Dead” (11:55–12:05, M5a), where he will show how treating prompts like code—versioned, tested, and continuously improved—delivers measurable ROI, reliability, and faster innovation. For technology and academic leaders, this session offers critical insight into moving from experimental AI to scalable, enterprise-ready generative systems. 🚀 #WorldSummitAI #EnterpriseAI #GenerativeAI #LLM #AIObservability #AIInnovation #AIProduction

Huge announcement, we've shipped Langflow 1.6, and it's packed with features for devs including OpenAI compat, native Docling, MCP + OAuth & more! 🚢 Let’s dive in 👇 1️⃣ 🔐 OAuth for MCP servers - Secure your MCP servers with OAuth, powered by MCP Composer. - Protect tools with authentication so only authorized users can access them. Walkthrough at: https://lnkd.in/gDByjjSn 2️⃣ 🤝 OpenAI Responses API compatibility - Works with OpenAI client libraries. - Makes integration dead simple. Docs: https://lnkd.in/gqdHmWq7 3️⃣ 📄 Advanced Doc Parsing with Docling - Parse PDFs & docs into clean structured data. - OCR support + VLM pipelines. Docling is now bundled—no extra install 🎉 4️⃣ 👀 Observability upgrade: Traceloop - Joins Arize, Langfuse, LangSmith & more to monitor agents in production. - Just set env vars and you’re live. Docs: https://lnkd.in/gbex5W69 5️⃣ 🎨 UX Improvements - Core components, bundles & MCP servers are organized - Quick search to find what you need fast A smoother building experience. 🚀 That’s Langflow 1.6! Highlights: 🔐 OAuth for MCP 🤝 OpenAI API compatibility 📄 Docling-powered parsing 👀 Traceloop observability 🎨 Better UX Read more about it here: https://lnkd.in/gtFscDkN

Another big win from WRITER Developer Experience for our customers: OpenLLMetry support is now available for the Python SDK via the Traceloop SDK! 🎉 Trace and observe your API calls alongside other LLM operations in your observability stack. Since it’s built on OpenTelemetry, you can use any OpenTelemetry compatible backend (Jaeger, Zipkin, Datadog, New Relic, etc.), evals platforms like Braintrust or LangChain's LangSmith, or the hosted Traceloop platform. Huge congrats to Yan Tolstoy and Ashley Weaver for building and documenting this and thank you Nir Gazit for your support! No more excuses: look at your data! It's super easy to get started. Check out these resources: 🛠️ OpenLLMetry repo: https://lnkd.in/gkg-Yu8N 📄 Integration guide: https://lnkd.in/gfsrj2jA

פלייבוק לדאשבורדים חכמים כדי לעקוב אחרי AI Agents ככל שהמערכות מבוססות AI הופכות ל-multi-agent ומנצלות MCP ו-Agent-to-Agent (A2A), האתגר האמיתי הוא לא רק לבנות – אלא לדעת למדוד, לפקח ולשפר. 📈 הנה דשבורד שהכנתי Base44 כדי לראות בעצמכם - https://lnkd.in/dBfbsAJB 🧭 אז מה חשוב למדוד כשבונים דאשבורדים ל-AI Agents? הנה הפלייבוק הבסיסי: 🎯 מדדים מרכזיים - דיוק ואיכות: שיעור הצלחת משימות, מדדי "goal completion", איתור הזיות, פידבק מהמשתמשים. - ביצועים ו-Latency: משך זמן לפנייה בודדת של Agent, זמן עד לתגובה ראשונה במצב סטרימינג, עומסים ותורים. - אמינות: אחוזי שגיאות, ביטולים, ניתוקים, יציבות אינטגרציות עם כלים חיצוניים. - עלות ויעילות: שימוש בטוקנים (prompt/completion), עלות לכל משימה/סשן, חיסכון בזכות caching. - הצלחת העברת משימות בין Agents, זמני handoff, נפח הודעות וארטיפקטים. - כמה פעמים ה-Agents מפעילים כלים זמני תגובה, אחוזי הצלחה/כשל. 📊 איך לבנות את הדאשבורדים? - מעקב אחרי MCP ו-A2A: תהליכי משימה, זמני הודעות, חיבורי סטרימינג. - צריכת טוקנים, עלויות בפועל מול יעדים. - Reliability – שגיאות לפי Agent/Model/Tool, עומסים ותורים. – הצלחת משימות, איכות תגובות, אינדיקציות לטוקסיות/דליפת מידע – יעילות כלכלית: עלות ל-Outcome, Utilization מול תקציב. 🔔 Alerts - ירידה באחוז ההצלחות של משימות - עלייה פתאומית בעלות למשימה - קפיצה בשיעור השגיאות או ביטולי ✅ מה יוצא מזה? - איתור תקלות מהר יותר עם Trace מקצה לקצה. - שליטה בעלויות ומניעת חריגות תקציב. - שיפור חוויית המשתמש והקטנת נטישה. - עמידה במדדי אמינות ובדרישות אבטחה. חברה שעוזרת לפתור את הבעיה ששווה לפרגן לה Traceloop #DataObservability #AIAgents #MCP #A2A #OpenTelemetry #Observability #MLOps #AIInfrastructure #CostOptimization #Latency #Reliability #base44

🚀 Traceloop is available in Google Cloud Platform! 🚀 I am thrilled to share that Traceloop is now part of the Google Cloud ecosystem! This partnership represents a major milestone  in our mission to make AI observability and LLM monitoring accessible to every developer and organization.  📺 https://lnkd.in/dwf7HPWR  #Traceloop #AI #GoogleCloud #LLMObservability

In this special episode of the ODSC Ai X Podcast, we dive into the real-world impact of GPT-5 – from routing and hallucination issues to cost savings and open-weight models. Joining us are two expert guests: 🗣️ Ivan Lee: Founder and CEO of Datasaur, who helps enterprises build private LLM stacks and has deep experience evaluating model upgrades. 🗣️ Nir Gazit: Co-founder and CEO of Traceloop, and co-creator of the OpenTelemetry Generative AI SIG, who brings insight into model routing, evaluation strategies, and observability tooling. Together, we unpack what GPT-5 actually changed – and what teams should do next. 🎧 Listen here: https://linktr.ee/odsc Key Topics Covered: 🔹Why GPT-5’s biggest shift is routing, not reasoning 🔹What casual vs. power users gained (or lost) with the rollout 🔹Hallucination benchmarks vs. real-world results 🔹Evaluation strategies using open-source tools like Phoenix and LangChain 🔹OpenAI’s OSS model release and its enterprise implications 🔹Why developers worry about black-box routing and lack of traceability 🔹How to migrate safely: pinning snapshots, running evals, shadow testing 🔹Whether GPT-5 gets us closer to AGI — or just better infrastructure 🔹What to expect from agent workflows, tool selection, and model specialization Memorable Outtakes: 💬 Ivan Lee: “GPT-5 is an upgrade for 98% of users—but for the power users, the loss of model choice felt like control was taken away.” 💬 Nir Gazit: “Of course every new model crushes it on benchmarks—they’re optimizing for the benchmarks. That doesn’t mean it works for your use case.” 💬 Ivan Lee: “OpenAI’s OSS release might be the bigger story than GPT-5. Suddenly, enterprises are back at the table.” References & Resources: - Datasaur: https://www.datasaur.ai - Traceloop: https://www.traceloop.com - OpenAI GPT-5  https://openai.com/gpt-5 - OpenTelemetry Project: https://opentelemetry.io - Traceloop OpenLLMetry: https://lnkd.in/dPHGiPNa - Phoenix (Arize AI open-source evals): https://lnkd.in/gFHmSyBB - LangChain Evals: https://lnkd.in/dEKKVnNF - GPT-OSS Open Weight Models by OpenAI: https://lnkd.in/dr6QNTAb - Claude + Model Context Protocol (Anthropic): https://lnkd.in/d6MuUNjD - ARC-AGI Leaderboard: https://lnkd.in/d5KVPgu8 🎧 Listen to the full episode here -  https://linktr.ee/odsc #GPT5 #ModelRouting #LLMEvaluation #OpenSourceAI #AIOperations #ObservabilityTools #AgentWorkflows #OSSModels #EnterpriseAI #ODSCpodcast

Accelerate your LLM development with Traceloop and Google Cloud #Traceloop #GoogleCloud #GoogleCloudSpringboard #GoogleCloudPartners