Shayan Hodai
Helping others to make data-driven decisions 🐇🎩
About Me
I have a computer engineering background with a postgraduate degree in AI. With over three years of experience, I thrive to make a positive impact wherever I can, either through analysis, building predictive models, or data pipelines. I especially like to leverage big data in financial, environmental, and healthcare applications. Schedule a 15 minute call
I was born in Tehran, Iran, and It has been more than a year since I came to Toronto, Canada, to follow my dreams in the fast-evolving technology industry.
When I was in high school, I used to examine all possible ways that a mathematical problem could be solved. Out of many trials and errors which most of them were unsuccessful, I would often find an interesting solution. This made me feel like I was exploring the unknown. I was part of the Astronomy Olympiad team of my school because ever since I can remember, I have looked at the sky and wondered what is really happening there and how it is even possible. Time and space are unimaginably huge! During school, I learned about mathematics, physics, and chemistry and their applications. They say it right: mathematics is indeed the language of the universe!
Coming forward, I entered the K.N.T.U., one of the top universities in Iran, to study computer engineering. I enjoyed, learned, and made valuable friends. I dived deep into programming and took optional courses in artificial intelligence. I knew that AI theories have been around humans for a century, but models could only be executed effectively a decade ago. With emerging GPUs and the massive amount of data generated via smartphones and servers daily, AI is not about theories anymore. It has become a real solution to solve problems even faster than humans.
I graduated in computer engineering, and knowing the importance of data in building machine learning models, I started working as a data engineer at a construction consulting company. The company had a massive amount of unstructured financial data stored in various sources, including digital files, Excel spreadsheets, etc. They were in no-use status because no one could analyze messy data. I was responsible for collecting, ingesting, processing, and serving them. By extracting, transforming, and loading (ETL processes) data, I migrated them to the data warehouse. Eventually, the company had the required data to gain knowledge through analysis and predictive models. This accomplishment made a significant positive impact on the company's revenue. The company was able to predict the most profitable time to start a construction project.
In 2022, I came to Toronto, Canada, to study Applied AI Solutions Development. Thanks to my professors, I learned a lot about statistical analysis and the best ways of presenting insights. I successfully graduated and made a great community of talented classmates with whom we share our ideas about the field.
Besides my professional journey, I play tennis, and even though my favourite player, Roger Federer, has retired, I still follow Grand Slam tournaments. I am also a big fan of Manchester United, Tin Tin, and Billy Wilder's cinema.
And I play electronic music! 🎢
I love connecting with new people to discuss common interests or try new things. If you find something that excites you on this website, please do not hesitate to contact me.
Area of Expertise
Analyzing credit card transactions and predict whether transactions are fraudulent using machine learning algorithms. The workflow includes data collection and exploration, data processing, feature correlation analysis, automated processing using pipelines, model building, performance evaluation through cross-validation, and fine-tuning the best-performing model based on precision, recall, and F1 score metrics.
Building a linear regression model to predict the number of visitors to a park based on various weather conditions. The dataset comprises 167 features, including different weather parameters recorded over time.
Skills: Python, R, SQL, NumPy, Pandas, SciPy, Scikit-learn, Matplotlib, Seaborn, Tableau, PowerBI, Jira
Skills: Python, SQL, Excel, ETL, Data pipelines, Stream processing, Apache Spark, PySpark, Databricks, Snowflake
Machine Learning Engineering
Building a machine learning software that scrapes tweets from a list of chosen accounts using snscrape, performs sentiment analysis on threads and replies using Hugging Face's BERT transformer, and creates and updates a database on MongoDB Cloud. The software also builds a REST API with six endpoints using Flask. The model was deployed on Azure.
Skills: Predictive models, Python, NumPy, Scikit-learn, PyTorch, TensorFlow, AWS, Azure
Generative AI (LLMs)
Integration of OpenAI's GPT-3.5 Turbo and Langchain framework for chat completion in a Streamlit production environment.
Skills: AI, NLP, LLM, Deep learning, Transformers, Hugging face, OpenAI, Llama, Langchain, Prompt engineering
And here is one of my sets 🌞