about 4 years ago
Did you know that participating in hackathons may result in job offers? We reached out to 2018 Power Up Automation winner, Russel Alfeche, to learn about his hackathon experience that led him to a job offer from UiPath.
What inspired you to develop your application, Classie-TicketClassifier?
Coming from a shared services organization, this is an idea that came up having experienced firsthand how teams are struggling to streamline service management processes, specifically incident ticket triaging.
Every large organization already has a responsibility matrix detailing which team handles what, reinforced by the historical data from their ticketing system. Why not leverage the hidden power within that data and let AI drive the rest, hence, this project.
What was the most difficult part of the development process? And how did you combat it?
The voiceover. In fact, I had to crowdsource it. If only I had a golden voice like my good friend Jonas Arguelles. I give thanks to him for sharing his voice for a noble cause as this. Kidding aside, If "generating a breakthrough idea" is to be considered part of the development process, it would be that. Then, data gathering and cleansing. For most machine learning projects, about 80% of the effort is spent on data cleansing. The main goals of this task are to identify missing data, remove errors/duplicates, and cross-out irrelevant features in order to create a reliable dataset while improving the quality of the training data for analytics. This process will ultimately drive more accurate decision-making.
In fact, even the most sophisticated and advanced machine learning functions go south and underperform when it is not supported with quality data (1) Large enough quantities that is suitable for the algorithm to be used (2) Clean, well prepared data. Based on my research here’s the recipe:
This shows that the only way to combat any difficulties for any project and turn it into a success story is to dedicate ample amount of time in research/exploration and have a continuous learning mindset.
What were the tools/languages you used to develop Classie-TicketClassifer?
I used VisualStudio with .NET 4.6.1 Framework to create my custom activities. In order to build and train my model, I used Azure Machine Learning services to provide the necessary infrastructure and resources. Finally, using UiPath Studio, several iterations of assembly and regression testing is poured into the project in order to ensure a quality output.
Do you have any tips you can give to developers when it comes to developing in UiPath studio?
First and foremost, follow established best practices. Make sure to design and organize your workflow including the architecture before even starting the build so that you will have a smoother development process.
What is your best advice for all hackers?
It will be to never stop learning and be resourceful. Leverage vast resources from UiPath enablement ecosystems. Starting with the academy you can deep dive on the RPA foundations and enable you to build your first automation project.
If your goal is to explore UiPath's possibilities, start building your first robot. If you’re stuck at some point where you think you cannot find the component that will serve your purpose, try UiPath GO!.
For more about Russel, check out this video:
If you have any questions about the hackathon, please post on the discussion forum.