TAATHEER RAMDIN
Kuala Lumpur,Malaysia
Summary
Works at the intersection of machine learning, AI, and operational systems, with a focus on reducing complexity without oversimplifying it. Experience has centered on designing solutions that impose structure on ambiguous processes, enabling more reliable and transparent decision-making. Particularly interested in problems where technical precision must coexist with practical constraints.
Overview
2
2
years of professional experience
8
8
years of post-secondary education
Work history
Full Stack Engineer Intern
Deloitte
Ebene, Mauritius
2025.01 - 2025.05
Insurance Automation Project
- Designed and implemented an automated health insurance claim processing pipeline using n8n for workflow orchestration
- Built modular backend services using FastAPI and Python for OCR processing, metadata validation, and fraud detection integration
- Integrated OCR capabilities using Microsoft Azure Computer Vision to extract structured data and bounding regions from medical documents and receipts
- Developed a decision engine to automate claim approval, rejection, and flagging based on extracted data and validation rules
- Implemented real-time workflow tracking and stage-wise processing visibility within n8n pipelines
- Automated cost estimation and payment trigger logic, reducing manual intervention in claims processing
- Designed RESTful APIs and integrated them with frontend components using HTML, CSS, and JavaScript
- Managed structured data storage and transaction handling using PostgreSQL
- Contributed to a scalable, production-oriented system architecture integrating backend services with workflow automation
Top-Listed Mauritian Insurance Company Database Project
- Designed and developed a relational database schema using Microsoft SQL Server for insurance-related data management
- Modelled entities including policyholders, claims, transactions, and coverage structures using ER diagrams and normalization techniques (up to 3NF)
- Implemented tables, constraints, primary/foreign keys, and indexing strategies to ensure data integrity and query efficiency
- Developed SQL queries and stored procedures for claim retrieval
policy tracking reporting and analytics - Optimised database performance through indexing and structured query design
- Ensured scalability and maintainability of the schema for real-world insurance system requirements
Accounting Clerk
Alffinanz Consulting Ltd
2022.11 - 2023.02
- Initiation into the professional world commencing as an Accounting Clerk, marking post A levels debut.
- This role afforded the opportunity to dive into the realm of finance and accounting.
- Collaborating seamlessly with seasoned accountants, while assisting in crafting monthly financial statements, encompassing balance sheets, income statements, and cash flow statements.
Education
Bachelor of Science - Computer Science, Data Analytics
Asia Pacific University of Techcology and Innovation
Malaysia 2023.03 - 2026.03
A Levels -
Royal College Curepipe
Curepipe, Mauritius 2020.01 - 2022.11
O Levels -
Royal College Curepipe
Curepipe, Mauritius 2018.01 - 2019.11
Skills
Programming & Development: Python, Java, C, C, JavaScript, Rust, MATLAB, HTML, CSS, FastAPI, Flask, Django, Nodejs
Databases: PostgreSQL, Microsoft SQL Server, MySQL, MongoDB, BigQuery
Machine Learning & AI: Scikit-learn, XGBoost, LightGBM, TensorFlow, Keras, PyTorch, Hugging Face Transformers, Google Vertex AI
Deep Learning: TensorFlow, Keras, PyTorch,
Computer Vision & Image Processing: OpenCV, PIL/Pillow, scikit-image, YOLO (v5, v7, v8), Tesseract OCR, EasyOCR, PaddleOCR
NLP & Text Processing: NLTK, spaCy, TextBlob, Hugging Face Transformers, BERT, GPT, RoBERTa
Document AI & OCR: Microsoft Azure Computer Vision, Google Vision API, Document AI, Tesseract, EasyOCR, PaddleOCR, LayoutLM, Donut (Document Understanding Transformer)
Workflow Automation: n8n
Tools, DevOps & Platforms: Git, GitHub, Docker, VS Code, IntelliJ IDEA, PyCharm, Eclipse, Postman, Linux, Ubuntu, Windows Server
Data Analytics & Visualisation: R, Pandas, NumPy, SciPy, Matplotlib, Seaborn, Plotly, Power BI,
Soft Skills: Leadership, Public Speaking, Critical Thinking, Problem Solving, Communication
Websites
LANGUAGES
English
Proficient (C2)
French - Native
Proficient (C2)
Creole
Proficient (C2)
Arabic - Intermediate
Intermediate
Extra Curriculars
Leadership & Communication
- Vice-President, APU Model United Nations (MUN) Club
- Treasurer, APU AI (AIC) Club
- Assistant Project Manager, APU MUN Debathon (2024)
- Committee Member, Asia Youth International Model United Nations (Kuala Lumpur)
- Committee Member, Island Night 2023 (Marketing, Design & Decor, Security, Social Media)
- Library Assistant, Asia Pacific University
Academic and Competitions
- Class Captain (Year 1) and Student Council Member (Year 2)
- Semi-finalist, Financial Services Commission (FSC) Young Talent Competition (2019)
- Australian Mathematics Competition (5 consecutive years): High Distinction (2018, 2019), Distinction (2017, 2020), Credit (2016)
- Alliance Française Programme (4 years): ‘Excellent’ (2014), ‘Très Bien’ (2015–2017)
Sports & Activities
- Royal College Curepipe Football Fiesta (3 participations): Gold Medal (2021), Silver Medal (2022), Bronze Medal (2019)
Projects
Multimodal AI Invoice Fraud Detection (Undergrad Final Year Project)
- Designed and implemented an end-to-end multimodal fraud detection system for digitised invoices, integrating image forensics, OCR, semantic NLP, and metadata analysis
- Built a structured preprocessing pipeline including PDF segmentation, high-resolution rendering, deskewing, grayscale conversion, denoising, CLAHE contrast enhancement, and adaptive thresholding
- Developed Agentic Document Extraction (ADE) pipeline using YOLO-based layout detection (YOLOv8) for region segmentation and spatially-aware key information extraction
- Implemented token-to-region assignment and fallback resolution using hybrid ML models (MultinomialNB, Logistic Regression, LinearSVC) with TF-IDF and contextual features
- Engineered synthetic fraud dataset with controlled multi-type tampering (text overlays, spatial manipulation, compression artefacts) for supervised training
- Built patch-based visual fraud detection models using EfficientNet-B0, ResNet50, and MobileNetV3, optimised for CPU execution with tf.data pipelines and hyperparameter tuning
- Applied image forensic techniques including Error Level Analysis (ELA), luminance inconsistencies, and compression artefact detection for tampering localisation
- Developed semantic fraud detection layer using rule-based consistency checks (VAT validation, duplicate detection, entity coherence) on OCR-extracted data
- Designed multimodal fusion framework combining visual, semantic, and metadata signals with deterministic scoring and confidence normalisation
- Implemented explainability layer using heatmap generation and interpretable scoring to highlight suspicious regions and justify model outputs
- Deployed system via local interface/dashboard enabling document upload, real-time analysis, and fraud visualisation
References
References available upon request.
Timeline
Full Stack Engineer Intern
Deloitte
2025.01 - 2025.05
Bachelor of Science - Computer Science, Data Analytics
Asia Pacific University of Techcology and Innovation
2023.03 - 2026.03
Accounting Clerk
Alffinanz Consulting Ltd
2022.11 - 2023.02
A Levels -
Royal College Curepipe
2020.01 - 2022.11
O Levels -
Royal College Curepipe
2018.01 - 2019.11