Web Application Design

Web Application Design

Empowering Insightful Decision Making

Empowering Insightful Decision Making

Empowering Insightful Decision Making

Hatari's ML Ops was designed to enchance insights for data transparency and Model Optimization.

Hatari's ML Ops was designed to enchance insights for data transparency and Model Optimization.

Hatari's ML Ops was designed to enchance insights for data transparency and Model Optimization.

Summary

Summary

01
Role

Product designer, product, product strategist and interaction designer

02
Overview

We conducted user interviews, surveys, and analyzed in-app analytics to understand the pain points and user needs. We also studied competitor apps and industry trends to gather insights

03
Challenge

We conducted user interviews, surveys, and analyzed in-app analytics to understand the pain points and user needs. We also studied competitor apps and industry trends to gather insights

04
What I did

We conducted user interviews, surveys, and analyzed in-app analytics to understand the pain points and user needs. We also studied competitor apps and industry trends to gather insights

05
Takeaway

We conducted user interviews, surveys, and analyzed in-app analytics to understand the pain points and user needs. We also studied competitor apps and industry trends to gather insights

01
Role

Product designer, product, product strategist and interaction designer

02
Overview

We conducted user interviews, surveys, and analyzed in-app analytics to understand the pain points and user needs. We also studied competitor apps and industry trends to gather insights

03
Challenge

We conducted user interviews, surveys, and analyzed in-app analytics to understand the pain points and user needs. We also studied competitor apps and industry trends to gather insights

04
What I did

We conducted user interviews, surveys, and analyzed in-app analytics to understand the pain points and user needs. We also studied competitor apps and industry trends to gather insights

05
Takeaway

We conducted user interviews, surveys, and analyzed in-app analytics to understand the pain points and user needs. We also studied competitor apps and industry trends to gather insights

01
Role

Product designer, product, product strategist and interaction designer

02
Overview

We conducted user interviews, surveys, and analyzed in-app analytics to understand the pain points and user needs. We also studied competitor apps and industry trends to gather insights

03
Challenge

We conducted user interviews, surveys, and analyzed in-app analytics to understand the pain points and user needs. We also studied competitor apps and industry trends to gather insights

04
What I did

We conducted user interviews, surveys, and analyzed in-app analytics to understand the pain points and user needs. We also studied competitor apps and industry trends to gather insights

05
Takeaway

We conducted user interviews, surveys, and analyzed in-app analytics to understand the pain points and user needs. We also studied competitor apps and industry trends to gather insights

01
Role

Product designer, product, product strategist and interaction designer

02
Overview

We conducted user interviews, surveys, and analyzed in-app analytics to understand the pain points and user needs. We also studied competitor apps and industry trends to gather insights

03
Challenge

We conducted user interviews, surveys, and analyzed in-app analytics to understand the pain points and user needs. We also studied competitor apps and industry trends to gather insights

04
What I did

We conducted user interviews, surveys, and analyzed in-app analytics to understand the pain points and user needs. We also studied competitor apps and industry trends to gather insights

05
Takeaway

We conducted user interviews, surveys, and analyzed in-app analytics to understand the pain points and user needs. We also studied competitor apps and industry trends to gather insights

Tools

Tools

Tools

Tools

Tools

Tools

The Design Process

The Design Process

Introduction

Introduction

The why

The why

The why

Take your love of food to the next level.

The ML Ops feature was added to Hatari to help users better handle and enhance their machine learning models. With this feature, users can see how transactions are scored, keep track of their models' performance, and quickly adjust them as necessary. It's like giving users a toolbox to fine-tune their models and stay ahead of the game. This way, users can make smarter decisions and adapt to any changes in Hatari with confidence.

The ML Ops feature was added to Hatari to help users better handle and enhance their machine learning models. With this feature, users can see how transactions are scored, keep track of their models' performance, and quickly adjust them as necessary. It's like giving users a toolbox to fine-tune their models and stay ahead of the game. This way, users can make smarter decisions and adapt to any changes in Hatari with confidence.

Target audience

Target audience

Understanding and identifying the users

Understanding and identifying the users

Understanding and identifying the users

Understanding and identifying the users

To create a useful ML Ops feature for Hatari, we needed to know who would use it. We talked to different kinds of users, like data scientists and developers, to understand what they needed. By learning about their work and what they struggled with, we made sure the feature would be helpful for everyone who uses Hatari to manage machine learning models.

To create a useful ML Ops feature for Hatari, we needed to know who would use it. We talked to different kinds of users, like data scientists and developers, to understand what they needed. By learning about their work and what they struggled with, we made sure the feature would be helpful for everyone who uses Hatari to manage machine learning models.

User research

User research

Talking to potential users

Talking to potential users

Talking to potential users

Talking to potential users

Before diving into designing the ML Ops feature for Hatari, it was important to talk directly to potential users. We reached out to different people who might use this feature, like data scientists and developers. By having conversations with them, we learned about their experiences, challenges, and what they wanted from a tool like Hatari. These discussions helped us gain valuable insights into their workflows, allowing us to design a feature that would truly meet their needs and make their jobs easier.

Before diving into designing the ML Ops feature for Hatari, it was important to talk directly to potential users. We reached out to different people who might use this feature, like data scientists and developers. By having conversations with them, we learned about their experiences, challenges, and what they wanted from a tool like Hatari. These discussions helped us gain valuable insights into their workflows, allowing us to design a feature that would truly meet their needs and make their jobs easier.

In the comprehensive user research step several valuable insights could be gathered. Here's some of what people had to saw:

In the comprehensive user research step several valuable insights could be gathered. Here's some of what people had to saw:

In the comprehensive user research step several valuable insights could be gathered. Here's some of what people had to saw:

In the comprehensive user research step several valuable insights could be gathered. Here's some of what people had to saw:

"Current tools lack integration capabilities with our existing infrastructure, causing data silos and inefficiencies."

"Current tools lack integration capabilities with our existing infrastructure, causing data silos and inefficiencies."

"Current tools lack integration capabilities with our existing infrastructure, causing data silos and inefficiencies."

"Current tools lack integration capabilities with our existing infrastructure, causing data silos and inefficiencies."

"Improving visualization tools could enhance our ability to interpret model outputs and make informed decisions."

"Improving visualization tools could enhance our ability to interpret model outputs and make informed decisions."

"Improving visualization tools could enhance our ability to interpret model outputs and make informed decisions."

"Improving visualization tools could enhance our ability to interpret model outputs and make informed decisions."

"Having a centralized dashboard for model monitoring and a streamlined process for model deployment are crucial for our workflow."

"Having a centralized dashboard for model monitoring and a streamlined process for model deployment are crucial for our workflow."

"Having a centralized dashboard for model monitoring and a streamlined process for model deployment are crucial for our workflow."

"Having a centralized dashboard for model monitoring and a streamlined process for model deployment are crucial for our workflow."

"A more efficient ML Ops solution could significantly reduce our time spent on manual tasks, allowing us to focus on model refinement and innovation."

"A more efficient ML Ops solution could significantly reduce our time spent on manual tasks, allowing us to focus on model refinement and innovation."

"A more efficient ML Ops solution could significantly reduce our time spent on manual tasks, allowing us to focus on model refinement and innovation."

"A more efficient ML Ops solution could significantly reduce our time spent on manual tasks, allowing us to focus on model refinement and innovation."

Brainstorming

Brainstorming

Bouncing ideas off each other

Bouncing ideas off each other

Bouncing ideas off each other

Bouncing ideas off each other

During brainstorming for the ML Ops feature in Hatari, our team collaborated to explore ideas such as improving model visualization tools and automating model retraining. We also discussed potential integrations with existing tools to streamline workflows. Through this process, we generated innovative concepts that laid the groundwork for further development.

During brainstorming for the ML Ops feature in Hatari, our team collaborated to explore ideas such as improving model visualization tools and automating model retraining. We also discussed potential integrations with existing tools to streamline workflows. Through this process, we generated innovative concepts that laid the groundwork for further development.

High Fidelity Design

High Fidelity Design

ML Ops for Hatari

ML Ops for Hatari

ML Ops for Hatari

ML Ops for Hatari

Our high fidelity prototype in Hatari is geared towards seamlessly integrating ML Ops functionalities into the platform. It begins with a user-friendly login process, followed by project creation and configuration, where users can upload data, connect streams, and train ML models.

Our high fidelity prototype in Hatari is geared towards seamlessly integrating ML Ops functionalities into the platform. It begins with a user-friendly login process, followed by project creation and configuration, where users can upload data, connect streams, and train ML models.

At the core of the design is the ML Panel tab, providing users with real-time insights into their ML model health. This ensures users can monitor model health effectively, with visual indicators prompting action if anomalies arise.

At the core of the design is the ML Panel tab, providing users with real-time insights into their ML model health. This ensures users can monitor model health effectively, with visual indicators prompting action if anomalies arise.

We delved into the realm of model retraining, offering users flexible options to configure retraining settings based on predefined drift levels. Users can opt for automatic retraining, where Hatari saves a fixed volume of data and triggers retraining upon detecting drift. Alternatively, users retain control with manual retraining, enabling them to define parameters such as data volume required, drift level threshold, and access a comprehensive log of model retraining activities.

We delved into the realm of model retraining, offering users flexible options to configure retraining settings based on predefined drift levels. Users can opt for automatic retraining, where Hatari saves a fixed volume of data and triggers retraining upon detecting drift. Alternatively, users retain control with manual retraining, enabling them to define parameters such as data volume required, drift level threshold, and access a comprehensive log of model retraining activities.

Takeaways

Takeaways

What I learnt from this

What I learnt from this

What I learnt from this

What I learnt from this

My takeaway from designing Hatari's ML OPS feature was the importance of user-centricity and collaboration in creating impactful solutions. By understanding user needs and working closely with diverse teams, I learned how to translate complex technical concepts into intuitive designs that enhance user experience and efficiency. This experience reinforced the value of iterative design processes and the significance of maintaining brand identity throughout product development. Overall, it highlighted the power of teamwork and empathy in driving successful product design.

My takeaway from designing Hatari's ML OPS feature was the importance of user-centricity and collaboration in creating impactful solutions. By understanding user needs and working closely with diverse teams, I learned how to translate complex technical concepts into intuitive designs that enhance user experience and efficiency. This experience reinforced the value of iterative design processes and the significance of maintaining brand identity throughout product development. Overall, it highlighted the power of teamwork and empathy in driving successful product design.