Liza Katz

Liza Katz

Seasoned Full-Stack Developer & GenAI Consultant | Ex-Elastic | RAG, Agents, LLMs, Search Optimization

About Me

I’m a seasoned software engineer and team lead with deep experience in building scalable, high-performance solutions for search, analytics, and GenAI applications. My background spans ad-tech, finance, and big data, working with technologies like JavaScript, TypeScript, Next.js, and Python.

I specialize in Retrieval-Augmented Generation (RAG), autonomous agents, hybrid and semantic search optimization, and consult on GenAI-powered products from architecture to deployment. As a consultant, I’ve delivered end-to-end projects for clients including Nayax, Cover, Max Security, and ScreenSteps.

Recent highlights include a Webby Award for the LLM game "Human or Not" and winning the Google AI Hackathon for "Tube Uni."

I’m always interested in new consulting projects and partnerships - if you’re building with GenAI, search, or LLMs, let’s connect.

Outside of work, I recharge through woodworking, travel, and baking.

Projects

**Webby Award Winner!** A Social Turing Game. Chat with someone for two minutes, and try to figure out if it was a fellow human or an AI bot. Think you can tell the difference?

Human or Not (AI21)

Implemented an agentic AI assistant for Max Security, a company that provides comprehensive protection and intelligence to proactively mitigate physical threats. The assistant has access to multiple geo political datasets and can provide question answering as well as advanced geo-spatial anlaysis of threats. Built on top of AWS, OpenSearch, and Claude.

Scout AI Agent (Max Security)

Cover is a company that provides insurance solutions. I helped the team design and implement data analytics pipelines and a RAG architecture to improve their claims processing and fraud detection capabilities. Built with AWS, OpenSearch, OpenAI and n8n.

Data Analytics Pipelines and RAG Architecture (Cover)

ScreenSteps is a company that provides a knowledge base and customer support platform. I helped the team develop a hybrid search solution that combines traditional keyword search with semantic search using ElasticSearch and AWS Bedrock. The solution allows customers to find relevant articles and answers to their questions more easily.

Hybrid Search for customer support (ScreenSteps)

**Google AI Hackathon Winner!** An AI-powered application that generates custom podcasts on any topic, using Gemini 1.5 Pro.

Tube Uni

Implemented a ReAct AI assistant for Nayax, a company that provides cashless payment solutions. The assistant is used to help vending machine operators maintain their machines and provide better customer service. I developed the assistant on top of a custom deployment of OpenAI on Azure, including prompt engineering, automation, evaluation and monitoring on top of Elasticsearch.

MoMa AI Assistant (Nayax)

Goods.io is a company that provites a single platform for order fullfulment and inventory management. I helped the team develop an analytics strategy, integrate Elasticsearch into their platform, and develop dashboards for their customers.

Goods.io - Elasticsearch integration and analytics

Outburn LTD is a company that provides a FHIR based API for healthcare providers. I helped Outburn ready their codebase for open source contributions, including an automation pipeline using GitHub Actions, proper usage of TypeScript and linting and contribution documentation.

Catomolo is a game that was developed for the Israeli environment protection agency. I helped the team update their infrastructure, including developing a modern next.js game server and setting up monitoring in place.

Contibution to the OpenSource project that allows running YOLOV5 inference directly in the browser.

Open Source implementation for a long requested plugin for Kibana allowing customers to Customize Kibana's appearance (logos, icons, texts and more)

Docker compose setup of an observability use case. Reporting metrics and traces from a nodejs app to Elasticsearch using APM Server using APM RUM and APM Node.js agents.

Organized 2 DjangoGirls conferences in the RedHat offices in Raanana

Experience

Gen AI Team Lead, August 2024 - Present
Leading a team of 3 engineers (Data Science, Full Stack and DevOps). Developed several production RAG applications on top of OpenSearch\ElasticSearch, using aws cloud tools such as Bedrock, SageMaker, EC2, AWS Lambda, ECS, S3 with all leading LLMs (Gemini, Claude, GPT-x)

Consultant & Full-Stack Developer, January 2023 - Present
Worked with several customers including AI21, mygoods.io, Outburn LTD and more. I focus on projects that focus on developing LLM-powered user experiences, powered by big data and search capabilities

Principal Full-Stack Developer, Performance Area Lead, January 2019 - December 2022
Worked on the Kibana architecture & services team, focused on delivering common services and React components to platform and solution teams. My contributions to Kibana, Kibana Load Testing, Kibana Capacity Test and more were made in private repositories. I was also the Performance Area Lead for Kibana, responsible for the performance of the Kibana platform.

Full-Stack Developer, Team Lead, May 2013 - April 2018
Developed the MobileCore SDK, a cross-platform SDK for mobile developers to monetize their apps. I was responsible for the SDK's architecture, design, and implementation. I then led a team of developers and QA engineers, to develop a mobile video advertising solution, supporting cutting edge ad formats, advertising standards and targeting options. A lot of my work was focused on ad serving business logic, which led to optimizations, cutting upto 70% of data and traffic costs.

Full-Stack Developer, 2008 - 2013
Worked as a fulltime freelance developer for a private equity firm. I was involved in multiple projects of different scales and complexities, using a variety of programming languages and environments, per customer requirements. Worked directly with customers from multiple industries, including finance, healthcare, and entertainment.

Embedded C \ C++ Developer, 2006 - 2008
AISeek developed an hardware that could accelerate the execution of A* algorithms on hardware, hence allowing to improve NPC AI in games. I was involved in the development of the hardware's firmware, and the software that interfaced with it. I also developed a custom C++ library for the hardware's API.

Skills

Python, (ES6+), TypeScript, Next.js, Node.js

Elasticsearch, OpenSearch, Redis, PostreSQL, BigQuery, RedShift

GCP, Vertex AI, AWS, Bedrock, Docker, Roboflow, Vercel, Github Actions