About Me

Hello Everyone! I am Zadan Fairuz Mahitala, an Applied Data Analyst / Junior Data Scientist with a strong emphasis on SQL-driven data architecture, large-scale data validation, and automated analytical workflows. I specialize in managing and validating nationwide survey data across 35 provinces, with proven ability to clean, join, and translate complex relational datasets into decision-ready insights.

My professional experience spans NLP-driven text classification, Python-based data processing, and practical machine learning applications for qualitative data standardization and electoral analysis. With additional proficiency in advanced Excel (formulas & VBA), I am committed to data quality, reproducibility, and translating analytical outputs into results that drive informed decisions.

Find my resume here!

Education

Degree	Institution	GPA	Relevant Coursework
B.Sc. in Informatics	Universitas Singaperbangsa Karawang	3.90/4.00	Data Structure & Algorithm, Statistic, Artificial Intelligence, Data Mining, Business Intelligence, Big Data Analysis, Machine Learning, Deep Learning, Digital Image Processing, Embedded Intelligent Systems

My Professional Journey

Data Science

| PT. Indekstat Konsultan Indonesia | October 2025 – January 2026

• Optimized data extraction, cleaning, and validation workflows for nationwide datasets spanning 35 provinces, achieving a 7% reduction in data errors and improving downstream reporting reliability.
• Leveraged advanced Excel techniques (INDEX-MATCH, VBA macros) and anomaly detection methods to identify and resolve discrepancies, ensuring accuracy for DPRD provincial legislative reports.
• Developed an NLP-based ensemble pipeline combining XLM-RoBERTa and IndoBERT to reclassify open-ended survey responses, standardizing qualitative data at scale for electoral analysis.
• Executed multi-source data integration (joining 2024 voter registries (DPT) with 2025 electability data) to build scalable legislative seat allocation models (Sainte-Laguë method), projecting seat distribution across 35+ regions and delivering actionable stakeholder insights.

Machine Learning Cohort

| Bangkit Academy led by Google, Tokopedia, Gojek, & Traveloka (Remote, Indonesia) | September 2024 – January 2025

• Selected as one of the top 10% of applicants from a pool of 45,000+ for an intensive 900-hour machine learning program, earning 12 industry certifications in NLP, TensorFlow, and Time Series analysis.
• Led a multidisciplinary team of 7 to develop “FindUp”, a talent-matching platform. Managed the end-to-end ML lifecycle including data acquisition, preprocessing (CV parsing), and model integration.
• Engineered a BERT-based regression model in TensorFlow to automate CV scoring (0–100 scale) and implemented a DistilBERT + K-Means clustering pipeline to map candidate skills to specific tech roles.

Data Analyst Student

| RevoU | August 2022 – December 2024

• Participated in MSIB 2023 - Kampus Merdeka program by the Ministry of Education, focusing on Data Analytics and Software Engineering.
• Software Engineering Courses: Mastered advanced HTML, CSS, and JavaScript; AJAX & Restful API; animations; transitions; and full-stack integration.
• Data Analytics Courses: Specialized in understanding business problems, EDA, SQL, Python for data analytics, BigQuery, and data visualization.
• Enhanced professional growth through structured career development sessions.

Legal & Licensing Assistant (Intern)

| PT Arkons Satya Konsultan, Bogor | January 2021 – May 2021

• Managed the end-to-end business licensing workflow via OSS (Online Single Submission) for 4+ clients, ensuring 100% regulatory compliance and on-time certificate issuance across multiple sectors.
• Spearheaded a long-term licensing initiative that improved client lifetime value (LTV) by 31% through streamlined administrative procedures and proactive cross-sector coordination.

Work Experience Highlights

• 🤝 Cross-functional Collaborator: Played a pivotal role in multidisciplinary teams, leveraging technical and analytical skills to bridge communication gaps between data scientists, engineers, and stakeholders for cohesive project execution.

• 💡 Balanced Decision-Maker: Known for balancing speed and quality in decision-making, particularly during high-pressure scenarios, ensuring timely delivery of data-driven solutions without compromising accuracy.

• 🌟 User-Centric Innovator: Spearheaded enhancements in user interfaces and functionality for machine learning models and analytics tools, resulting in a 25% increase in user engagement and streamlined workflows.

• 🔍 Data-Driven Strategist: Conducted in-depth analyses and delivered actionable insights through compelling dashboards and reports, enabling stakeholders to make informed business decisions aligned with organizational goals.

• 🧠 Problem-Solving Pioneer: Tackled data challenges by designing automated data pipelines and implementing optimized machine learning algorithms, reducing processing time by 40% and improving model reliability.

Projects

FindUp: Bridging Tech Talent and Startups

A capstone project developed as part of Bangkit Academy 2024. This mobile platform connects tech talents with startup founders, addressing challenges in collaboration and team dynamics.

• Integrated TensorFlow, Keras, and a BERT-based model for enhanced talent matching.
• Achieved 89% accuracy in resume analysis and recommendation systems using NLP techniques.

Challenge: Build a platform to connect tech talent with startup founders by automating CV analysis, scoring, and role matching, addressing inefficiencies in recruitment and enabling personalized recommendations.

Action:

• Designed a BERT-based regression model to score CVs (0–100) based on skills, certifications, degrees, and work experience, combining embeddings with dense layers for accurate predictions.
• Implemented an automated CV parsing pipeline to convert PDFs into structured JSON data for seamless model input.
• Built an unsupervised text similarity model to align candidates with startup roles such as Data Scientist or Product Manager.
• Collaborated with backend engineers to integrate models into the app, ensuring real-time inference using TensorFlow.js and preprocessing with Google Cloud Functions.
• Optimized data pipelines and trained models on scalable cloud infrastructure for reliability and performance.

Result:

• Reduced recruitment time by streamlining candidate prioritization and role matching, enabling startup founders to focus on top talent.
• Enhanced the platform’s personalization features, improving user engagement and satisfaction.
• Contributed to scaling Indonesia’s startup ecosystem by enabling AI-driven hiring solutions aligned with national digital transformation goals.

Technologies:

1. NLP Models: Pre-trained BERT, TensorFlow
2. Data Processing: JSON structuring, tf.data API
3. Cloud Services: Google Cloud Run, Pub/Sub, Cloud Functions, Vertex AI, JWT
4. Mobile Stack: Retrofit, Datastore, Gson, JUnit & Espresso, Fragment, Logging, Navigation, CircleImageViewglide,Lifecycle and ViewModel

Documentation: Github Repo

Real-Time Sign Language Detection

Challenge:
Develop a system that utilizes digital image processing and machine learning to recognize and interpret sign language gestures in real-time, aiding individuals with hearing impairments in inclusive communication.

Action:

• Utilized TensorFlow Object Detection API to create a real-time sign language detection model.
• Collected 15 images for each of 5 sign language gestures: “hello,” “yes,” “no,” “thank you,” and “I love you.”
• Implemented a 6-step process:
  1. Captured real-time images using a webcam and labeled gestures.
  2. Converted datasets into TFRecord format for training and testing.
  3. Applied transfer learning with a pre-trained SSD MobileNet V2 FPNLite model for fine-tuning.
  4. Configured a model pipeline and trained it over 10,000 steps to enhance accuracy.
  5. Restored the model checkpoint for real-time inference.
  6. Developed functionality for real-time gesture detection via webcam.

Result:

• Achieved a high accuracy rate in detecting and interpreting sign language gestures.
• Enhanced accessibility for individuals with hearing impairments by providing a scalable and inclusive communication tool.
• Identified areas for future improvements, such as adapting to variations in sign language and improving model accuracy, to further support accessibility in technology.

Customer Cancellation Analysis for Resort Hotels

Provided actionable insights to reduce cancellation rates and optimize revenue strategies for hotels.

Challenge: Analyze and reduce hotel booking cancellation rates from 13–29% to 10% within a year while addressing revenue decline caused by customer cancellations.

Action:

• Data Cleaning: Processed raw datasets by handling null values, correcting data types, addressing outliers, and removing logical errors using Python.
• Exploratory Data Analysis (EDA): Performed linear regression and data analysis with Python and Google Sheets to understand trends and patterns.
• Visualization: Built insightful dashboards using Looker Studio to highlight trends in booking cancellations, ADR (average daily rate), and system performance.
• Root Cause Identification: Conducted an issue tree analysis to pinpoint factors contributing to cancellations, such as system delays, deposit policies, and lack of service updates.
• Recommendations & Policies: Suggested key improvements, including deposit requirements, improved system response times, and better customer communication.

Result:

• Identified high cancellation rates for bookings without deposits (49.29% for Resort Hotels and 49.14% for City Hotels in 2017).
• Proposed implementing deposit policies to lower cancellation rates.
• Provided strategies to optimize revenue by reducing lead times, increasing staff, and upgrading systems to improve service quality.
• Highlighted actionable insights for improving ADR competitiveness and minimizing cancellation impact.

Additional Insights & Recommendations:

• City Hotels can attract more customers by lowering prices by at least 9.4% and enhancing service quality.
• Resort Hotels should focus on promoting leisure travel to capitalize on their lower cancellation rates.
• Sending booking reminders and offering rescheduling options can reduce cancellations and enhance customer retention.

Google Colab Link: Data Cleaning

Skills & Certificates

Programming Languages: Python, SQL, VBA
Databases & Cloud: MySQL, Google BigQuery, Neo4j (Graph DB)
AI & Machine Learning: TensorFlow, Scikit-Learn, NLP, Computer Vision, Deep Learning, Information Retrieval
Data Analysis & BI: Looker Studio, Power BI, Advanced Excel/Spreadsheet, Power Query, EDA
DevOps & Deployment: Docker, FastAPI, Redis

Certificates I Gain: here

🔗 Connect with me