• Developed trading strategies in Python, yielding positive returns for clients by leveraging proprietary linear regression techniques on economic and stock features. • Deployed trading models into GCR and their corresponding features in to Google Storage Buckets. • Applied quantitative analysis and created back-testing model to assess the strength and risk of each feature in trading strategies.• Strengthen strategy performance through continuous data research and developing innovative quantitative techniques, leveraging research papers and team brainstorming.• Built automated data ingestion pipeline using FactSet, Bloomberg, web scraping and other financial vendor APIs, totally over 500K raw records, into Postgres database. • Contributed to financial LLM training and enhancement by creating and evaluating over 500 prompts, including testing codes in SQL, Python, and question prompts, ensuring accurate and concise responses.• Created an automated daily trade execution tool using the Interactive Broker Python SDK and trade signals from GCR, streamlining the trading process.• Implemented an automated daily update email for clients, featuring trading signal, bar-charts of trading features KPIs, and a strategy performance table. Used Python, HTML, Google Email.

• Developed stock recommendation model using linear-regression and a forecast model to assess whether users’ portfolios outperform or underperform the S&P500 benchmark. Used Python.• Applied quantitative analysis methods to stock and cryptocurrency to extract key insights, such as performance trends and correlations, to improve both the recommendation and forecast models. • Built and maintained data pipelines to collect 10K+ stock and cryptocurrency data from Alpaca, Refinitiv, Polygon, Messari, and web sources into Azure Data Studio for ML models and data analysis.

• Built an algorithm with Folium and Scipy to map water sample locations and show water quality health level (1-safe to 5-avoid) contour to the user.• Used Time-Series model on a given precipitation measurement to forecast water quality levels for 2 days into the future. • Collaborated with UX/UI designer to optimize webpage interface for users.

• Built an algorithm with OpenCV to detect traffic lanes in a video • Applied Gaussian Blur, Canny function and Hough Transformation to identify lanes.

12 Weeks, 480+ hours, of Data Science training program.(Deep Learning, machine learning, statistic modeling and analysis, E.T.L)Capstone Project: Car Classification with Imagery (Auto-enVision) ** - Built and deployed a CNN car image classification model with 98% accuracy, leveraging a TensorFlow, into a web application. Pretrain Model. Use Python, Pandas, NumPy, Flask, HTML and Heroku.- Built a data pipeline to gather images, processed images and stored them in a local folder to be access by the machine learning model. Used Python, BeautifulSoup, Pandas, NumPy, PIL and PyPlot. Client Project: Flood Detection with Imagery (Flood Bank)-In a group of 3, developed a flood classification algorithm and flood detection with pretrained model to filter flooded images and then use it approximate its flood depth. -Built a web scraper tool to gather flooded images. -Used LabelImg for image labeling and used Roboflow’s Yolo-v3 Pytorch pretrained model for flood detection.-The model achieved accuracy of about 70% with 1000 images and demonstrated the capabilities of the concept. -The leading group who thought outside the box and went beyond project requirements.Housing Price Prediction (Houstradamus) -Created a Linear Regression algorithm to predict housing prices with the given features of the house.-Performed exploratory data analysis to extract import features and data cleaning of the data. -Incorporated feature engineering and regularization to minimize mean-square-error.-Achieved 2nd in the leaderboard within my class and 4th within General Assembly Data Science classes.

• Improved shipping delays by 80% from developing a data tracker to forecast customer needs and initiating production ahead of schedule. Used Excel and Exact Macola 9.• Optimized 10+ assembly BOMs by more than 30% and reduced production cost by 50%. Analyzed/revised part designs and worked with assemblers to identify areas of overdesign and material redundancy. Used Solidworks, 3D printer, Excel, and Exact Macola 9.

-Successfully implemented a harness safety cover into driver footwell from concept into production by working with drivers to understand their driving discomforts and movements. -Executed a customer order of over 10 buses for worth over $500K+ by working with the customer to custom fit a special camera to the exterior wheel panel.

-Researched and developed calibration mounts for helicopters swashplate assembly for wind-tunnel testing. -Designed hardware capable for wind-tunnel testing by doing FEA simulation against the maximum vibration estimated by analyzing the wind-tunnel dataset.-Collaborated with team members and machinists to improve, optimize and reduce the cost of the calibration mount

-In a team of four to research, design and develop a suture device for a heart surgeon.-Performed FEA on the suture device using Solidwork to withstand an impact of 4pounds. -Cooperated with the team and client to optimize the final design for ergonomic and material.-Performed drop test and penetration test on suture device to ensure the rigidity of the device.

-Computer designed a motorcycle chassis capable of two 200lbs battery packs, a 300lbs electric motor, two axles and a brake system for each wheel to convert a motorcycle to a 3 wheeler.-Simulated and optimized suspension setup with spring stiffness and damping characteristics to achieve maximum vertical acceleration of 0.058G for best comfort.