M. Thufail Alwannabil Samas

Data & AI Professional

M. Thufail Alwannabil Samas

Practical data systems, machine-learning solutions, and AI automation.

Experience spanning data analysis, forecasting, model evaluation, Python/PostgreSQL pipelines, API services, and LLM-assisted workflows with human review.

View Work Download Resume GitHub LinkedIn
Surabaya, Indonesia Career focus: Data, Machine Learning, and Applied AI

About

Data Analysis, Machine Learning, and AI Systems

Professional experience covers operational monitoring, data science, and technical education. Work at PT Aplikanusa Lintasarta centered on DWDM monitoring and assurance analytics for approximately 400 directional sensors. Earlier roles involved BMKG forecasting tools and programming and AI education across school and university settings.

Beyond those roles, the portfolio includes machine-learning forecasting and classification, data pipelines, and LLM-assisted invoice automation with human review. Credentials include BNSP Data Scientist certification and a first-author peer-reviewed publication on XGBoost rainfall forecasting.

Selected Work

Applied Data & AI Case Studies

Start with the problem, contribution, and outcome. Open the technical details when you want to inspect the implementation and validation choices.

Filter by focus

Showing all seven projects.

01

AI automation & document intelligence | n8n, Gemini, Python/FastAPI, PostgreSQL, Docker

End-to-End AI Invoice Approval Automation

Problem
Finance teams lose time and introduce avoidable errors when invoices arrive in different layouts and must be re-entered before totals, references, duplicates, and approval rules can be checked.
My Contribution
Designed and implemented the workflow end to end: multi-file intake, three n8n automations, Gemini extraction, a Python/FastAPI validation service, PostgreSQL workflow history, human review, and duplicate-safe delivery to an ERP sandbox.
Technical details
Data & Context
The interface accepts PDF, PNG, and JPEG vendor invoices in batches of up to 20 files. Each document keeps its own source identity, processing state, and audit history.
Approach
n8n creates one job for every uploaded file, starts extraction and validation, then routes the result to approval, finance review, or an unreadable-document queue. One failed invoice does not stop the rest of the batch.
Validation & Controls
Python rules check file integrity, required fields, totals, currencies, references, dates, duplicates, and approval routing. Docker Compose packages the local services, while an optional Apps Script integration sends final read-only records to Google Sheets.
AI invoice automation flow showing batch intake, automated checks, human review, accounting delivery, and a structured record of status, decisions, and duplicate protection
One project is split into three n8n workflows so intake, human review, and accounting delivery can start independently and be retried safely.
20 filesMaximum Upload Batch
3Event-Driven n8n Workflows
Per-fileFailure Isolation

Project OutputFinance users can upload a batch, review only the values that need attention, authorize ready invoices, and deliver approved records without retyping each document. Failed files remain isolated instead of blocking the full batch.

Evidence & ScopeThe runnable local project includes structured invoice records, exception queues, approval tasks, searchable history, audit events, and duplicate-safe ERP sandbox records across n8n, Gemini, Python/FastAPI, PostgreSQL, and Docker Compose.

View repository
03

Time-series forecasting | Python, XGBoost, scikit-learn

Rainfall Forecasting with XGBoost and Grid Search

Problem
Daily rainfall is highly variable, and random train-test splits can overstate how well a model will perform on future weather. This project estimates next-day rainfall while keeping evaluation in chronological order.
My Contribution
Built the Python forecasting workflow from data validation and lag construction through XGBoost search, scikit-learn temporal evaluation, and experiment reporting, then served as first author of the resulting peer-reviewed publication.
Technical details
Data & Context
Daily BMKG observations from the Trunojoyo observation area in Sumenep Regency combine recent temperature, sunshine, humidity, wind, and rainfall history.
Approach
Created one-day lagged predictors and searched 243 XGBoost configurations, with Min-Max scaling fitted separately inside each training fold.
Validation & Controls
Used 10-fold TimeSeriesSplit and pooled out-of-fold predictions so every test window followed its training data and future observations did not leak into preprocessing.
Observed rainfall and pooled out-of-fold XGBoost predictions across ten temporal test windows
Observed rainfall and pooled out-of-fold predictions from the 10-fold temporal evaluation.
0.9152MAAPE
11.9566RMSE (mm)
10Temporal Folds

Project OutputThe workflow delivers a selected XGBoost model, fold-level results, pooled out-of-fold rainfall predictions, and a traceable experiment report rather than a single unsupported accuracy figure.

Evidence & ScopeThe work became a first-author peer-reviewed publication. Original BMKG observations remain restricted, so the public repository provides synthetic same-schema data for reviewing and reproducing the workflow.

View repository Read publication
04

Clinical classification | Python, XGBoost, scikit-learn

Thyroid Cancer Recurrence Classification with XGBoost, Information Gain, and Grid Search

Problem
Build a reproducible recurrence classifier while testing whether a smaller set of informative features can preserve performance and reduce unnecessary model input.
My Contribution
Implemented the Python data audit, scikit-learn preprocessing, fold-local Information Gain selection, checkpointed XGBoost search, and multi-metric evaluation for this collaborative academic project.
Technical details
Data & Context
The public UCI dataset contains 383 differentiated thyroid cancer records with clinical, pathological, treatment, response, and recurrence variables.
Approach
Handled numeric, binary, ordinal, and nominal features explicitly, then combined entropy-based Information Gain selection with XGBoost and Grid Search.
Validation & Controls
Repeated feature selection inside every training fold and evaluated pooled out-of-fold predictions with 10-fold cross-validation, preventing the full dataset from influencing feature choice.
Top ten encoded features using their original column names, ranked by mean Information Gain across ten training folds
Top 10 encoded features by mean Information Gain across 10 training folds.
97.25%Accuracy
95.37%F1-score
98.05%Specificity
98.60%ROC-AUC

Project OutputThe workflow delivers fold-specific feature rankings, pooled out-of-fold predictions, and a reproducible comparison of XGBoost configurations across accuracy, F1-score, specificity, and ROC-AUC.

Evidence & ScopeThe public UCI data and runnable implementation are available in the repository. Results are strong within this experimental dataset, but the project supports model research rather than clinical diagnosis.

View repository
05

Data engineering & monitoring analytics | Python, PostgreSQL, HTML/CSS/JavaScript

DWDM Optical Sensor Monitoring with Change-Point and Degradation Detection

Problem
Operations teams need a consistent way to separate missing telemetry, short failures, persistent signal shifts, and gradual degradation across hundreds of directional optical sensors.
My Contribution
Implemented the incremental Python/PostgreSQL pipeline, data-quality and failure rules, change-point and degradation analysis, paired-sensor evidence, local API, priority workflow, and browser-based monitoring interface as a Junior Programmer at PT Aplikanusa Lintasarta.
Technical details
Data & Context
Approximately 400 directional sensors combine native 5-minute PRTG optical-power telemetry with ENIMS metadata for sensors, links, directions, endpoints, and locations.
Approach
Built an incremental workflow from metadata synchronization and telemetry collection through failure grouping, PELT change-point detection, Theil-Sen and Mann-Kendall degradation analysis, paired-sensor evidence, priority ranking, and local dashboard delivery.
Validation & Controls
Statistical analysis uses only clean observations. Missing data, downtime, zero coverage, and low-signal failures remain visible to operations but are excluded from change and trend calculations instead of being filled with invented values.
Illustrative DWDM optical signal trace with a change point, degradation trend, failure interval, missing interval, and monitoring output
Illustrative, non-confidential view of signal events, missing intervals, and monitoring output used by the workflow.
~400Directional Sensors
5 minTelemetry Cadence
3Operational Views

Project OutputThe pipeline turns raw telemetry into prioritized, traceable investigation views: failure events, persistent downward shifts, gradual degradation, paired-sensor evidence, and dedicated operations, event, and management screens.

Evidence & ScopeThe analysis helps teams decide what to investigate first but does not claim a physical root cause. Internal telemetry, source code, credentials, infrastructure, database contents, and operational screenshots remain confidential.

06

Deep learning & time-series forecasting | Python, TensorFlow/Keras, NumPy, Pandas

Rainfall Forecasting with BiLSTM and Grid Search

Problem
Next-day rainfall depends on changing weather patterns across several recent days, making a sequence-based model more appropriate than treating each observation as an isolated row.
My Contribution
Developed the Python forecasting workflow during my BMKG internship, using NumPy and Pandas for seven-day sequence preparation and TensorFlow/Keras for three-layer BiLSTM training, focused Grid Search, chronological evaluation, and result reporting.
Technical details
Data & Context
Daily BMKG observations from the Trunojoyo observation area in Sumenep Regency provide four weather variables and historical rainfall for each input sequence.
Approach
Built seven-day multivariate sequences, fitted Min-Max normalization on the training period, and trained a three-layer stacked BiLSTM while tuning units, batch size, and the learning-rate drop period.
Validation & Controls
Used an 80:20 chronological split so every evaluation observation followed the training period. Predictions were returned to millimeters and impossible negative rainfall values were clipped to zero.
Observed rainfall and next-day BiLSTM predictions over the chronological test period
Observed rainfall and next-day predictions over the chronological test period.
0.8073MAAPE
10.2734 mmRMSE
7 daysInput Sequence

Project OutputThe workflow turns seven days of weather history into next-day rainfall estimates and preserves the selected model, holdout predictions, evaluation summaries, and actual-versus-predicted visual for review.

Evidence & ScopeThe reported BMKG internship results come from one chronological holdout period using restricted observations. The public repository includes synthetic same-schema data so the full workflow can still be inspected and run.

View repository
07

Statistical time-series forecasting | Python, statsmodels, Pandas

Rice Price Forecasting with ARIMA and Walk-Forward Validation

Problem
Price planning benefits from a forecast that is easy to explain and tested in the same order that future monthly prices would actually become available.
My Contribution
Implemented the Python/statsmodels ARIMA workflow for this collaborative academic project, covering Pandas-based monthly preparation, stationarity checks, order selection, diagnostic screening, expanding-window evaluation, and forecast reporting.
Technical details
Data & Context
Sixty public monthly observations from January 2020 through December 2024 cover medium and premium rice prices from the Open Data Jawa Timur portal.
Approach
Modeled medium and premium prices separately with ARIMA, using ADF-based differencing, ACF and PACF inspection, candidate evaluation, convergence checks, and residual diagnostics.
Validation & Controls
Started with the first 80% of observations, forecast the next month, added the newly available actual price, and repeated the process across the remaining 12 months.
Observed and one-step-ahead forecast prices for medium and premium rice during walk-forward evaluation
Observed and one-step-ahead forecasts over the walk-forward test period.
2.08%Medium Rice MAPE
1.82%Premium Rice MAPE
12 monthsWalk-Forward Test

Project OutputThe workflow turns public monthly price history into explainable one-step-ahead forecasts, selected ARIMA orders, diagnostic records, error summaries, and comparison visuals for both rice categories.

Evidence & ScopeThe public data and runnable workflow are included in the repository. The models tracked the evaluated period closely, but rely only on past prices and do not include market events, policy changes, supply, or weather.

View repository

Publication

Peer-Reviewed Rainfall Forecasting Study

First Author | Peer-Reviewed Article

Prediksi Curah Hujan di Kabupaten Sumenep Menggunakan Metode Extreme Gradient Boosting (XGBoost) dan Algoritma Grid Search

M. Thufail Alwannabil Samas, Nurissaidah Ulinnuha, Moh. Hafiyusholeh

MIND (Multimedia Artificial Intelligent Networking Database) Journal, Vol. 11, No. 1, June 2026, pp. 30–43

Open DOI

Experience

Professional Experience

Junior Programmer

PT Aplikanusa Lintasarta

Developed a project-stage DWDM optical monitoring and assurance analytics system for approximately 400 directional sensors, integrating 5-minute PRTG telemetry with ENIMS network metadata through an incremental Python/PostgreSQL pipeline, failure detection, statistical change-point and degradation analysis, paired-sensor evidence, and dashboard-based operational monitoring.

Coding and AI Teacher

SMP GIKI 3 Surabaya

Taught Python fundamentals and introductory AI concepts to Grade 8 and 9 students over five months, using Teachable Machine and practical coding exercises while designing lesson materials, quizzes, and image-classification activities to strengthen programming logic, problem-solving, and foundational AI understanding.

Data Science Intern

Stasiun Meteorologi Maritim Kelas II Tanjung Perak Surabaya (BMKG)

Developed an internal extreme-weather monitoring dashboard, processed and compared BMKG cloud-prediction imagery with Himawari IR observations using Python, and built a 3-layer BiLSTM rainfall forecasting model using 7-day sequences and an 80:20 chronological split, achieving MAAPE of 0.8073 and RMSE of 10.2734 mm.

Programming Teaching Assistant

Universitas Islam Negeri Sunan Ampel Surabaya (UINSA)

Guided undergraduate students across four semesters of Basic Programming and Programming courses in MATLAB and Python, providing hands-on support in debugging, data handling, computational problem-solving, and algorithmic thinking while developing introductory machine-learning materials covering Linear Regression, K-Nearest Neighbors, and K-Means clustering.

Skills

Capabilities Across Data and Applied AI

Data Analysis & Statistics

Python, SQL, R, Pandas, NumPy, data preparation, exploratory analysis, statistical modeling, and visualization

Machine Learning & Evaluation

Scikit-learn, XGBoost, Statsmodels, TensorFlow/Keras, feature engineering, cross-validation, and time-series evaluation

Data Engineering & APIs

PostgreSQL, MySQL, ingestion, transformation, validation, incremental pipelines, FastAPI, and REST APIs

AI Automation & Integration

n8n, Gemini API, structured LLM output, document extraction, webhooks, human review, and duplicate-safe workflows

Development & Delivery

Git, GitHub, Docker, unittest, Jupyter Notebook, HTML, CSS, JavaScript, and dashboard development

Education

Bachelor of Mathematics (S.Mat.)

Universitas Islam Negeri Sunan Ampel Surabaya (UINSA)

August 2021 – January 2025 | GPA: 3.46/4.00 | Completed the 4-year bachelor's program in 7 semesters (3.5 years)

Certification

Data Scientist

Badan Nasional Sertifikasi Profesi (BNSP)

Valid September 2024 – September 2027

Contact

Data & AI Opportunities

I am open to Data Scientist, Data Analyst, Data Engineer, Machine Learning Engineer, AI Engineer, and AI Automation opportunities, including remote roles.

mthufailalwannabils@gmail.com
Surabaya, East Java, Indonesia
LinkedIn GitHub Resume