Complete gig history
This talk discusses evaluating and securing LLM applications by measuring changes in prompts or RAG pipelines. It highlights evaluation frameworks like Vertex AI Evaluation, DeepEval, and Promptfoo, and introduces security measures using LLM Guard to ensure resilience against prompt injections and harmful responses, emphasizing the need for robust input-output guardrails.
It’s easy to generate content with a Large Language Model (LLM), but the output often suffers from hallucinations (fake content), outdated information (not based on the latest data), reliance on public data only (no private data), and a lack of citations back to original sources. Not ideal for real-world applications. In this talk, we’ll provide a quick overview of the latest advancements in multi-modal LLMs, highlighting their capabilities and limitations. We’ll then explore various techniques to overcome common LLM pitfalls, including Retrieval-Augmented Generation (RAG) to enhance prompts with relevant data, ReACT prompting to guide LLMs in verbalizing their reasoning, Function Calling to grant LLMs access to external APIs, and Grounding to link LLM outputs to verifiable information sources, and more.
WebAssembly (Wasm) allows you to compile code written in over 40 programming languages and run it in a secure and performant way in web browsers. The WebAssembly System Interface (WASI) has expanded the capabilities of Wasm by enabling it to run outside the web browser, such as server-side applications, edge computing, and cloud microservices. Docker has also recently announced support for Wasm, allowing it to be used as a lightweight alternative to Linux and Windows containers. Whether Wasm will replace containers remains to be seen but it’s definitely worth learning more about it. In this talk, I’ll introduce Wasm, the basic terminology around it, and its current state as a server side technology. We will also look at some examples and tools for working with Wasm on the server side.
AsyncAPI is an open-source specification for describing and documenting asynchronous APIs, similar to OpenAPI specification for documenting RESTful APIs. CloudEvents is a specification for event data in the cloud. Together, they enable developers to design, document, and test event-driven APIs and to easily share and consume event data across different cloud platforms and ecosystems. In this session, we will explore the benefits of using AsyncAPI and CloudEvents in your tech stack, and how they can help you build asynchronous, event-driven applications that are well-documented and easy to maintain. METE ATAMEL I’m a Software Engineer and a Developer Advocate with 18+ years of experience. Currently, I work in the Developer Relations team at Google in London. I build tools, demos, tutorials, and give talks to educate and help developers to be successful on Google Cloud. As a regular speaker at tech conferences, I have spoken over 345+ events since 2016 on modern application development topics such as Kubernetes, Istio, Knative, serverless, event-driven architectures, and microservice orchestration. Prior to my current role, I was a Software Engineer/Architect/Tech Lead at Nokia, EMC, Adobe, Skype, and Microsoft building various client and server technologies.
We went from a single monolith to a set of microservices that are small, lightweight, and easy to implement. Microservices enable reusability, make it easier to change and scale apps on demand but they also introduce new problems. How do microservices interact with each other toward a common goal? How do you figure out what went wrong when a business process composed of several microservices fails? Should there be a central orchestrator controlling all interactions between services or should each service work independently, in a loosely coupled way, and only interact through shared events? In this talk, we’ll explore the Choreography vs Orchestration question and see demos of some of the tools that can help.