Jose Antonio Pepe Tapia

Resume

Bloom Chat

Generative AI on Enterprise Data. Increase productivity where it matters.

Uniting different data sources to improve productivity

Enterprise data is often siloed and difficult to access. By creating a chat interface that can consolidate historic and real-time data into one interface, we can help users discover insights and make better decisions.

Creating a full stack application

The goal was to create an application that can consolidate historic and real-time data into one interface. By utilizing modern NLP technologies, we can help users discover insights and make better decisions.

Design, Develop, Deliver

I took the initial design and narrative documents and decided to focus heavily on the backend infrastructure and AI implementation. By utilizing a data centric approach, I was able to integrate multiple data sources such as Microsoft 365, Quickbooks, Phonecall transcripts, and a ticket system to deliver a working beta.

Results

After a successful beta, the user can now increase their productivity where it matters. Writing executive reports down from 3 days to 10 minutes. Meeting times down from 45 minutes to 5 minutes. Response times from emails and phone calls down from 1-3 days to 4 hours or less.

Applied AI Full Stack Developer

As an Applied AI Full Stack Developer, I was responsible for the entire software development lifecycle. From initial design to final deployment, I was able to deliver a working beta in 2 months. By utilizing modern NLP technologies, I was able to help users discover insights and make better decisions.

Iterate. Fail. Learn. Repeat.

I'm not a UI/UX designer.

I am a problem solver. I took the product concepts and learned bits of Color Theory, UI/UX, and other design principles to create usable interfaces, gathered user feedback, and iterated from there.

As long as the functionality was there, it was more important to iterate, fail, and learn than to just not do anything at all.

Orange Crew Version 1

Create React App in 2023?

Learning best practice and modern tooling.

I did not use any UI component libraries for this project. I did everything in React, and when interactions were needed, I used JavaScript and CSS. Later, I transitioned to using Tailwind CSS for faster development. I used Axios to make API calls to connect to my FastAPI backend.

When learning, I find the documentation and examples because I promote self-sufficiency. If I get stuck on something, I find experts or peers to learn from.

Here are my favorite places to refer to when I was working on this project.

You might not need an effectDescribing the UI
Orange Crew Chat Interface

Do not reinvent the wheel.

Simplicity is key. But so is understanding the fundamentals.

I initially tried to impress rather than being practical. I focused on getting the core functionality working first and then iterated on the design later. Just because I am learning something, it doesn't mean I can't enforce high standards. While implementing the design, I constantly referred to best practices for React fundamentals.

Orange Crew Landing Page Version 5

Get the job done.

A new UI/UX designer has entered the chat.

When the product reached a stage that required a more polished look, the client hired a UI/UX designer to take over the design. I worked closely with the designer to implement the new design and ensured that the functionality remained intact. For data privacy reasons, here is the mock-up design that was implemented, with only a few compromises made in the final implementation.

Bloom Chat rebrand dashboard.