Wojciech Kulikowski

Client

The client was a company specializing in robotic process automation. For this project, NU4IT needed support with web development, marketing, and Facebook.

Project Description

From the beginning, the project had a clearly specified goal—to develop a web app that allows Facebook marketers to rapidly experiment with their Facebook Ads parameters.

The users should be able to:

- Login with Facebook and log out
- See and review their ad campaigns, ad sets, and ads
- Create a new campaign
- Create a config that defines multiple ad sets at once, publishes them, and allows them to monitor which ones perform the best
- Create an ad for those ad sets and run the experiment
- Monitor experiment

Outcome

We delivered a working MVP after one month of development. Every user story has been successfully delivered, and the app has successfully launched to over 200 users.

In MVP, all ads are image ads, and users can define their title, message, image, and call to action.

After the whole campaign has been created, users can analyze the performance in the ad tracking tab.

Technology used

- Python
- Django
- Facebook Marketing API

A KYC AML solution to integrate with fintech solutions. Authenticating users are going through the following steps:
1. Basic information input
2. Face scan/liveness check
3. Front and back ID scan
4. Face match from photos and the video
5. Check public watchlists

The back end was developed with Django and deployed with Kubernetes on AWS.

For a German client, we built a complex recruitment tool. A recruiter could generate a job posting and a chatbot that performs the initial candidate screening. Then, with only one click, the posting is being deployed as a targeted ad on Facebook, alongside the corresponding messenger chatbot.
I was responsible for building the chatbot system. It has been developed as a separate service and put under the biggest load when the app went to production.

A marketplace connecting homeowners with renovators. The back end was developed in Django and exposed as a REST API to a React front end. We hosted the solution on AWS ECS, an Amazon container orchestration tool.

Publicist is a marketplace connecting PR specialists and companies in need of PR services.
We built the project with Django, React, and Stripe. I was responsible for developing the back end specifically for the project management flow and payments.

The system had four responsibilities:
- It should collect historical data back to 2018
- It should run in hourly intervals to collect up-to-date orders from Amazon and Shopify
- It should run in hourly intervals to collect up-to-date ad spending from Facebook, Google, and Amazon
- It should calculate the profit margins for every product (with regard to platform commissions, taxes, and shipping costs)

The project has been hand over within a month. It correctly collected all the historical data, running for over 24 hours due to Amazon & Shopify rate limits.

From this point, the data collection process runs as an hourly cron job.

Challenges:
- After completing the project using Amazon MWS, was released. Amazon Selling Partner API would be more effective but it had not yet been released.
- It was difficult to acquire Shopify fee data. Finally, we succeeded by looking at the payment receipt info. It was not documented well and we had to dig into many pages of stack overflow and Shopify community answers.

Technology used:
- Python
- Google Sheets API
- Amazon MWS
- Shopify API
- Facebook Marketing API
- Google Ads API
- Amazon Advertising API