Google Vertex AI – Nobl9 Case Study

How F33 Contributed to the Development of Nobl9's Reliability AI Using Google's Vertex AI

About our client

Nobl9 helps software developers, DevOps practitioners, and reliability engineers deliver reliable features faster.

Through software-defined Service Level Objectives (SLOs) they can help link monitoring and other logging and tracing data to user happiness and business KPIs.

Client's challenge


Nobl9 aimed to develop a tool that would streamline and simplify the reliability-checking process, replacing the need for dashboards and the manual analysis of multiple indicators.


This tool would function as a Virtual Assistant, primarily capable of addressing three types of questions: inquiries about data, questions related to the Nobl9 company, and general machine learning queries.


This led to the necessity of creating and deploying a website that would enable users to log in and utilize the Virtual Assistant.


As well as implementing an authentication mechanism to ensure customers can securely add and manage their keys

Introducing the solution


Utilizing Google’s App Engine and Google’s Vertex AI to develop a Virtual Assistant capable of addressing three main types of questions: queries related to customer data (e.g., identifying unreliable services), questions about Nobl9 (e.g., details about Nobl9), and general machine learning inquiries.


Assisting in the creation and deployment of a website that enables users to log in and utilize the virtual assistant, along with establishing an authentication mechanism using Google Cloud Key Management Service to ensure customers can securely add their keys.


Deploying all the components to make the tool accessible to the public, resulting in an openly available tool for anyone at, along with a corresponding blog post on the Google website.

Project results

AI-driven SLO Management

F33 played a crucial role in developing Nobl9’s AI interface, This interface allows users to upload a graph image, generating an SLO based on the presented line. It also supports inquiries about the resulting SLO, Error Budget, and recommended next steps.

Positive impact of implementation

Users can access wanted information quickly and easily, without a need to manually search through their dashboards. They can ask questions in natural language and get detailed, accurate answers about the state of their services and use the VA to filter and transform the information into something valuable.

Differences from the previous state

Prior to this, users had to manually access their dashboards and review all indicators on their own. There was no convenient method for retrieving relevant information using natural language in a practical manner.

Take the first step and

Google Cloud Platform services used

Google App Engine

Usage: Served as a scalable web server for hosting the application, providing a hassle-free development experience.

Benefits: Abstracted away infrastructure concerns, allowing a focus on code development, while ensuring automatic scaling based on demand.

Google Cloud Key Management Service

Usage: Served as a reliable tool for securely storing keys and managing expiration dates.

Benefits: A secure solution for storing and managing keys, eliminating the need for creating your infrastructure.

Google Vertex AI
Model Garden - PaLM

Usage: Implemented as a Large Language Model (LLM) from the Model Garden to intelligently answer user questions.

Benefits: Leveraged pre-trained language models for natural language processing, enhancing the intelligence and responsiveness of the application.

Google Vertex AI
Search & Conversation

Usage: Integrated as a context retriever to provide relevant information and enhance the search functionality.

Benefits: Empowered the application with robust search capabilities, enabling the retrieval of contextually pertinent data for an enriched user experience.


A word from our client

“What a team! Our friends at F33 crushed it building this #GenAI product! Highly recommend working with them!”

Kit Merker, Chief Growth Officer


F33 has successfully implemented and delivered a Proof of Concept for a chatbot based on LLM, effectively combining the strengths of Enterprise Search and PaLM 2. This accomplished solution caters to the customer’s specific needs by infusing intelligence into data through the utilization of Google Cloud Platform services. The implemented chatbot now stands as a testament to F33’s commitment to delivering innovative and efficient solutions, providing users with an enhanced experience in navigating and extracting valuable insights from their data.

The project’s success underscores a positive shift towards leveraging advanced technologies, highlighting the potential for continued innovation and positive outcomes in the customer’s data-driven endeavors.


Throughout the project, we gained valuable insights into the customer’s data quality and provided recommendations for maintaining high-quality data. This is essential for ensuring a seamless flow of information within the company.


Our efforts aimed to make the client aware of both the possibilities and limitations of technology, expanding their perspective on potential future avenues.

F33 team working on the project

User Icon Vector Male Person Symbol Profile Circle Avatar Sign in Flat Color Glyph Pictogram illustration

Ireneusz Wieczorek

Senior Backend Developer

Cloud Architect Certificate

machine learning engineer

Adam Czyżewski

ML Engineer

ML Certificate

machine learning engineer

Wit Jakuczun

PhD, Principal Data Scientist

ML Certificate

machine learning engineer

Maria Naklicka

ML Engineer

ML Certificate

Consult us about
your needs

During the consultation, we will carefully examine your situation and advise you on planning a solution.

Free of charge.

You can expect:

30-minute consultation

One-on-one with an AI expert

Exclusively focused on your business cases and challenges

Fill in the form


Our specialist will contact you in 24h