Built a Machine Learning model to predict the logistics cost based on the data of 16 logistics partner of the
client from last 3 years. This was to optimise the best possible options to minimise the transit time and tariff for the
freight.

Requirement of Machine Learning model to do analysis and prediction using last 9 years data of iron-ore mines of
a particular segment of the client and do a performance prediction. Client requirement was to do segment analysis
of all of their 4 segments in Chile. The model was built using Logistic Regression. The resulting
model had F1 score of 0.942 . It resulted in a 12% reduction in Freight cost with annual savings of $3.3M .

To build a content-based recommendation system using customer feedbacks and ASIN numbers. The ASIN
number is a textual information of products. The model was developed by pre-processing textual information
using NLP techniques and then finding similarity of textual information using Minkowski ā€™s distance. The final
model was having RMSE of 0.289. The resulting model increased user engagement by 32%.