I Am Salim BenHamadi

The Hackathon was a great opportunity for me to tackle the stock market domain, as I was required to implement a dynamic optimization algorithm for portfolio allocation, according to the Sharpe ratio. And to build a  time series forecasting algorithm to predict the stock prices of the optimized portfolio.

It is important to understand that a portfolio is no more than just a collection of stock investments held by an investor. Optimizing a financial portfolio requires selecting the best possible portfolio out of a set of all portfolios being considered according to an objective.

According to modern portfolio theory, one possible way to optimize a portfolio is to aim for an efficient frontier. The efficient frontier is an investment portfolio that lies on the most efficient(ideal) coordinates of the risk-return graph(spectrum). There is a formal definition that is used by many experts in this field, "Efficient frontier is the set of portfolios which satisfy the condition that no other portfolio exists with a higher expected return but with the same standard deviation of return.

Now we know the aim is to achieve one of the efficient frontier portfolios. However, we can even go one step further than the overall portfolio. We can identify the best possible proportion of the stocks to use in any given portfolio. This is possible by using the Sharpe ratio.

The Sharpe ratio was developed by William Sharpe in 1966. The Sharpe ratio describes how much excess return of money one receives for the extra volatile time he/she holds onto the investment. It measures how much extra capital money one can receive by waiting and holding onto a risky stock investment. The Sharpe ratio can be calculated by taking the difference between the returns of the investment and the risk-free return, divided by the standard deviation of the investment.

Used technologies :  Pypfopt, Keras, LSTM, Seaborn, Matplotlib, Numpy, Pandas, 

My Notebooks : 

• Diversification notebook : https://www.kaggle.com/code/salimhammadi07/value-diversification
• Forecasting notebook : https://www.kaggle.com/code/salimhammadi07/value-forecasting

- Worked collaboratively with a team to develop an eCommerce website. 

- Maintained quality assurance through code testing and cross-browser testing. 

- Troubleshot and debugged code ensuring compatibility with devices and browsers. 

- Participated in planning, developing, and testing a web application.

- HTML5/CSS

                    -Typescript

- React

- Bootstrap

- Redux

- MongoDB

- NodeJS

- ExpressJS

- Git/Gitlab

- Jest unit testing 

 My internship report :

With the growth of world population, much more effort and innovation will be urgently needed in order to sustainably increase agricultural production, improve the global supply chain, decrease food waste, and ensure access to nutritious food to all vulnerable people essentially in developing countries and Africa. Sustainable agriculture is related to The Sustainable Development Goal to “End hunger, achieve food security and improve nutrition and promote sustainable agriculture” (SDG2).

In this context, crop type mapping is a major challenge for agricultural and environmental policy makers. The recent growth of open Satellite imagery time series allows large scale crop mapping.

The objective of the challenge is to use Sentinel-2 multispectral time series to classify crops in Central Tunisia, specifically in the Kairouan agricultural region.

The ground truth reference data was collected in the field by the AGEOS Team. The challenge datasets were processed by IEEE Sup'com GRSS Chapter Members.

I am happy to announce that I placed the 3rd out 20 participants in the Landslide Prevention and Innovation Challenge by START Hack 2022.

Even though, my participation for this hackathon was unofficial, I Had a great time competing  with the other participants. also, It was a great opportunity to  enhance my knowledge in the subject of this competition.

About the competition : 

The common practice of identifying landslides is visual interpretation which, however, is labor-intensive and time-consuming. Thus, this hack will focus on automating the landslide identification process using artificial intelligence techniques, and target at using high-resolution terrain information to perform the terrain-based landslide identification.

The objective of this hackathon is to classify if a landslide occurred or not.

Link of the competition : https://zindi.africa/competitions/start-hack-22

Link of my notebook : https://www.kaggle.com/code/salimhammadi07/landslide-prevention-and-innovation-challenge

I am happy to share with you all that I won the 2nd prize 🥈 at the ML Olympiad during the Machine learning March which was hosted by Google Developer Student Club, ENIT, IEEE ENIT Student Branch & its IEEE Computer Society-ENIT Student Chapter, and the IEEE ESSTHS Student Branch and its IEEE ESSTHS CIS Chapter in collaboration with Google Developers.

📌 The objective of this challenge is to build a classification model that answers the question: “what culture does this image represent? using a dataset, composed of raw images of architecture, food, clothing, tattoos, festivals, statues, monuments, and paintings from north-African countries ( Algeria, Egypt, Lybia, Morocco, Mauritania, Sudan, and Tunisia).

Karim Omrane, Sahar Tekaya and I placed 2nd out of 52 active participants competing in this hackathon.