You will get Custom ML Model for Predictive Analytics
Top Rated
Project details
With over 4 years of hands-on ML expertise as a Software Engineering senior (joint UCD program), I stand out by delivering end-to-end pipelines that solve real-world problems in healthcare, 3D vision, and predictive analytics. What sets my projects apart is my proficiency in gradient boosting (XGBoost/CatBoost), tabular deep learning (TabNet/TabPFN), causal inference (PEHE/ATE), and optimization techniques like feature engineering and model calibration. Clients get robust, interpretable models with high accuracy, backed by Kaggle Gold/Silver medals and patent-pending innovations. I ensure scalable, error-free solutions that boost efficiency and decision-making—let's turn your data into actionable insights!
Machine Learning Tools
BERT, ChatGPT, Keras, MATLAB, NumPy, OpenCV, pandas, Python, Python Scikit-Learn, PyTorch, scikit-learn, SciPy, SQL, Stata, TensorFlow, XGBoostWhat's included
| Service Tiers |
Starter
$500
|
Standard
$800
|
Advanced
$1,000
|
|---|---|---|---|
| Delivery Time | 5 days | 7 days | 10 days |
Number of Revisions | 1 | 1 | 1 |
Number of Model Variations | 0 | 1 | 0 |
Number of Scenarios | 1 | 1 | 1 |
Number of Graphs/Charts | 10 | 10 | 10 |
Model Validation/Testing | |||
Model Documentation | |||
Data Source Connectivity | |||
Source Code |
5 reviews
(5)
(0)
(0)
(0)
(0)
This project doesn't have any reviews.
CS
Chinmay S.
Apr 29, 2026
Looking for someone to build an AI/automation to find crime articles (RecentlyBooked)
JK
Justin K.
Apr 25, 2026
AI / Python Engineer – Build Automated Video Analysis Pipeline (1000+ Videos)
QL
Q L.
Mar 23, 2026
Trading System Engineer – Market Regime & Structure Scanner (Step 1 Only)
LR
Lukas R.
Mar 14, 2026
AI automation Expert
Great work, thank you.
DG
Deb G.
Jan 21, 2026
Protein Folding Rate Analysis -- Model Comparison Study
Liu Yichen is a rare find on Upwork: a true scientist who understands the physics behind the data, not just the code.
He delivered:
1. Immaculate Code: Fully documented, reproducible Python scripts.
2. Publication-Ready Rigor: A formal PDF audit report that is journal ready.
3. Scientific Insight: He didn't just run the numbers; he interpreted them and improved our methodology.
If you need high-level bioinformatics, statistical modeling, or Python data science, hire him immediately. He is a research partner of the highest caliber.
He delivered:
1. Immaculate Code: Fully documented, reproducible Python scripts.
2. Publication-Ready Rigor: A formal PDF audit report that is journal ready.
3. Scientific Insight: He didn't just run the numbers; he interpreted them and improved our methodology.
If you need high-level bioinformatics, statistical modeling, or Python data science, hire him immediately. He is a research partner of the highest caliber.
About Liu
Machine Learning Expert
100%
Job Success
Beijing, China - 4:06 pm local time
Most of my work revolves around tabular data: gradient boosting (LightGBM, XGBoost, CatBoost), tabular deep learning (TabNet, TabPFN), and causal inference. I've spent a lot of time on the unglamorous but critical parts — feature engineering, handling imbalanced datasets, model calibration, and making sure results are interpretable and reproducible.
A few things I've shipped:
A cancer risk prediction system for a medical research team (AUC > 0.84), with full feature importance analysis and threshold strategy reports.
A 3D reconstruction pipeline based on Structure-from-Motion, robust enough to handle blurry and low-texture inputs. That work led to 2 software copyrights and a patent under review.
A causal boosting framework (CBDT) for heterogeneous treatment effect estimation — benchmarked on IHDP, ACIC, and MIMIC-III. Paper currently under submission.
Quantitative models for financial risk — credit scoring with XGBoost and TabNet, A/B testing on simulated datasets, ended up improving accuracy by 15% and cutting false positives by 20%. I'm comfortable working with financial data, backtesting pipelines, and building risk models from scratch.
I also work with OpenClaw for robotic manipulation tasks and have experience integrating it into simulation and control workflows.
On the tools side: Python is my daily driver, along with PyTorch, Transformers, Optuna, and Git. I write clean, documented code — not the kind you need to reverse-engineer to understand.
I'm finishing up a B.Eng. in Software Engineering (Beijing University of Technology Ă— UCD joint program, class of 2026). Kaggle Gold and Silver medalist.
If any of this sounds relevant to what you're working on, drop me a message. I usually reply within a few hours.
Steps for completing your project
After purchasing the project, send requirements so Liu can start the project.
Delivery time starts when Liu receives requirements from you.
Liu works on your project following the steps below.
Revisions may occur after the delivery date.
Carefully read the requirements document and the source code
Fully understand the customer's needs, and then fulfill the customer's requirements with the least amount of changes.