Econometrics | Deep Learning | Time Series Analysis
Manh Ho Duc
Aspiring Quantitative Analyst & Financial Data Scientist building model-driven systems for market research, causal inference, and production-grade forecasting.
MODEL MONITOR
Quant research dashboard
GPA
3.83/4
Actions
108
Focus
VNIndex
ABOUT
Career Objective
A data-science profile shaped for quantitative finance, combining mathematical foundations with economics, machine learning, and applied research.
Highly motivated and analytical Data Science student with a strong foundation in Mathematics and Economics, dedicated to pursuing a specialized career in Quantitative Finance. Short-term goal: secure a position as a Quantitative Analyst or Financial Data Scientist. Long-term aspiration: become a Quantitative Researcher or Lead Data Scientist.
National Economics University (NEU)
Data Science in Econometrics & Business
GPA: 3.83/4
High School for Gifted Students
Vinh University
SKILLS & INFRASTRUCTURE
Technical Operating Stack
A practical mix of financial modeling, machine learning, deep learning, visualization, and deployment capabilities.
Machine Learning
Deep & Reinforcement Learning
Frameworks & Libraries
Data Visualization
Cloud Compute & Deployment
QUANTITATIVE PROJECTS
Research and Model Portfolio
Interactive project cards focused on quantitative methods, econometric inference, forecasting, and deep learning.
Research
Dynamic Spillover Effects Analysis
Applied MS-VAR with Bayesian MCMC and GFEVD models to analyze spillover effects between Gold prices, Oil prices, and the VNIndex.
Compared empirical behavior with global benchmarks including S&P 500 and Shanghai Index.
Econometrics
Green Innovation & Financial Performance
Built econometric models for listed firms in Vietnam to evaluate how green innovation connects to financial performance.
Used OLS, Two-Way Fixed Effects, and GMM for causal inference under panel-data constraints.
Machine Learning
Stack Ensemble Weather Forecasting
Developed a robust prediction system using Gradient Boosting and Random Forest on historical weather data.
Added automated retraining triggers to respond to model drift and maintain forecast quality.
Deep Learning
Video Action Classification
Fine-tuned a Video Masked Autoencoder with a ViT backbone to classify 108 distinct human actions.
Implemented layer-wise learning rate decay and extracted embeddings for data-drift monitoring.
CONTACT
Discuss Quant, Data, or Research Roles
Use the form for a quick note, or reach out directly by email or phone.