Make Pets Found (using AI) - Day 2

4 minute read

Day 2

Everyone was in agreement that they’d had a flipping good night’s sleep. It must be the Andalusian air (or the absence of screaming toddlers, for the dads in the group). One of the many joys of El Molino Del Conde is the fresh produce available literally on the doorstep. Simon convinced Matt and Hugo to put their city boy ways on hold and collect eggs from the chicken coop in time for breakfast. Matt’s insistence on ‘taking the chicken from behind’ was an inspired decision, and we left the coop with a basket of freshly laid eggs.

Fully fed, we cracked on. Henry and Sam worked on getting image recognition models up and running in AWS SageMaker — this allows us to create, train, and deploy machine learning models for real time inferencing, all in the cloud.

Meanwhile, after wrapping up some basic CRUD for the Serverless backend, George started to integrate our app with the endpoints provided by SageMaker, using our trained model. Luke ploughed on with normalising the response we received from Google Cloud Vision, into a format we can store and parse.

We were ticking along nicely, but still our WiFi woes continued. Up steps la protagonista, Jesús. After putting in some calls, he jumped in the car and headed south to Antequera to pick up what was required. Not only did Jesús return with internet, he also arrived laden with cases of Alhambra. Not all heroes wear capes.

Laura stepped up to the plate (or wok) for our evening meal, cooking up a stir fry that went down an absolute storm. After dinner saw a collaborative effort between devs and… not devs, as Jesús, Sam and Matt all helped Hugo with a video export that was just not playing ball. Eventually, the team drifted off to bed (following some late-night Rocket League from the usual suspects) and the cow jumped over the moon.

Dia dos, en la bolsa.