About
Growing up in Hawassa, Ethiopia, my passion for gaming sparked an early fascination with technology, which has since evolved into a focused interest in AI. Currently, I am part of the RAIL Lab, where I focus on safe reinforcement learning, aiming to contribute positively to advancements in the field and the intersection of technology and society.
Work Experience
Tech Stack
Latest research projects
Join me as I explore the latest trends, practical applications, and innovative techniques in the rapidly evolving field of AI.
Feedback-driven Safe Offline RL
Integrating expert feedback into offline RL to enhance decision-making in safety-critical environments while minimizing risks associated with live trials.
Learning WVFs without Exploration
Exploring how offline RL algorithms learn a goal-oriented representations - World Value Funnctions (WVF), emphasizing the importance of dataset size and diversity.
I like building things
From developing trading bots and implementing deep RL algorithms to creating insightful data visualizations, each project highlights innovative solutions and practical applications.
- L
Learning World Value Functions without Exploration
Researched and implemented offline RL algorithms, including DQN and BCQ, to learn WVFs across various data buffer sizes in discrete and continuous action domains.PythonPyTorchNumPyOpenAI GymPanda Gym - M
ML trading bot with Alpaca API
Developed an ML based trading bot using the Alpaca API to automate stock trading decisions based on historical market data analysis.PythonAlpaca APIPandasNumPyScikit-learnMatplotlib - W
Web scraping summarizer
Developed a web scraping summarizer that extracts key information from web pages and generates concise summaries for easier content consumption.PythonScrapyNLTKNumPyPandas - D
Deep RL for Franka Emika Robot
Implemented Deep Deterministic Policy Gradient (DDPG) to enable the Franka Emika Panda robot to learn and perform complex tasks in dynamic environments.PythonPyTorchOpenAI GymNumPyPanda Gym - C
Credit card fraud detection
Developed a POC for credit card fraud detection system using machine learning algorithms to identify and prevent fraudulent transactions in real time.PythonScikit-learnPandasNumPyTensorFlow
Get in Touch
If you're into AI, data science, and occasional tech memes, we should definitely connect on LinkedIn!