The Traceway eve aimed at transferring experience and networking between programmers and senior developers in the field of artificial ielligence and machine learning was organized by Quera on December 14 at the Energy Faculty of Sharif University. At the beginning of the program, Mr. Tabrizi, CEO of Quera, gave a brief explanation about Quera startup and how to run the eve.
In the first part of the preseations,Abbas Hosseini Technical Vice Preside of Tapsell explained about the three advertising software of Pegah company named MediaAd, Tapsell and Tagrow, how to display an advertiseme related to each user by predicting the behavior of users in fro of the various advertisemes they see.
Second preseation by Ali Chalmaqaniartificial ielligence product manager at Cafe Bazar, and explained about natural language processing at Cafe Bazar (which has three software, Bazar, Divar, and Beled).
He stated that due to the importance of data and its processing in the curre era and the number of one billion ActionLogs collected from users, about 40 to 50 data scieists are working in Cafe Bazaar.
A number of issues with text processing in Cafe Bazaar have been resolved so far. One of the solutions is chat wall, which is investigated and with the help of Unsupervised ML, the conversations were grouped io about 200 clusters and its topics were tagged. By using these clusters, disturbances caused in chats are ideified and preveed.
The third preseation Nice message Data Scieist in Beled Router did a traffic story in Beled. He stated that the suggested routes are generated by receiving manual reports and also receiving GPS signals from users. These signals are placed on the correct routes displayed on the map using Map Matching.
One of the challenges is to estimate the speed of users when they eer the tunnel, and it is obtained by calculating the average speed at the beginning of eering the tunnel and the speed when leaving the tunnel.
The last preseation by the last preseation by Hamed Dehghani The data engineer from DigiKala was about recommender systems, which is now in DigiKala. Previously, third party systems were used, but now engineers have created an optimized system, leading to lower costs and better results.
At the end of the eve, a short preseation was made by the startups Salam Cinema, Vakavik, and Virgul, and each of them gave a brief explanation of how they operate.
Networking was done in such a way that an address was recorded on the back of each person’s card, and people with similar signs sat at a table and answered a number of questions in order to get to know each other better.
How do you rate this article?
