Profile of mohammad haziq mohd fadhil
mohammad haziq mohd fadhil senior machine learning engineer Mohammad Haziq
+6016-9065378 | haziq.fadhil92@gmail.com
Notice Period: 2 Weeks
Kuala Lumpur, Malaysia
Professional Summary
Experienced and results-oriented Senior AI Engineer with 5+ years of hands-on expertise building
intelligent systems across computer vision, natural language processing (NLP), and machine
learning. Proven ability to lead full-cycle AI product development — from ideation and dataset
design to deployment and post-deployment performance monitoring. Skilled in designing scalable
APIs, LLM-driven chatbots, recommender systems, OCR pipelines, and eKYC solutions for
industries such as finance, automotive, and telecommunications. Passionate about applying AI to
solve real-world problems and optimize business processes.
Skills & Expertise
• Machine Learning & AI: LLMs (OpenAI, Ollama, LlamaIndex), NLP, Recommendation
Systems, Deep Learning, Image Processing, OCR, Face Recognition, Object Detection,
Image Segmentation, Agentic AI, Vision LLM, Function Calling, RAG.
• Tools & Frameworks: TensorFlow, PyTorch, PaddlePaddle, Scikit-learn, OpenCV, SpaCy,
FastAPI, Flask, Docker, Android Studio, Alembic, Pinecone, AWS Lambda, AWS Api
Gateway, AWS Bedrock.
• Tech Stack: Python, REST APIs, SQL, Git, Postman, tensorflow, torch, huggingface.
• Deployment & Ops: Docker, GPU Setup, Data Pipeline Automation, Model Serving
• Soft Skills: Project Leadership, Technical Communication, Agile Development, Cross-
functional CollaborationProfessional Experience
Senior Software Developer / AI Engineer
Raizon Consulting Sdn Bhd — Kuala Lumpur, Malaysia (Hybrid)
July 2025 – Present
Leading AI initiatives the focused on applications of LLM for multiple purposes such as document
Intelligence. Responsible for managing project roadmaps and timeline for the Chatbots.
Key Contributions:
• Multipurpose Applications Chatbot
◦ Leading the optimization to reduce the chatbot slow response.
◦ Implementing the streaming response by creating a websocket api gateway to the
lambda function in AWS.
◦ Involved in creating product roadmap to improve user experience.
◦ Involved in revamping the Chatbot workflow by implementing the agentic
workflow .
◦ Involved in revamping the knowledge base in AWS Bedrock. Improving the
chunking strategies and vector storage.
◦ Database design for LLM system prompt storage.
• Video Analytics
◦ Leading the creation of POC for Video Categorization
◦ Created POC with using zero shot classification method.
◦ Includes the Multiclass classification
◦ Added the audio classification.
Senior Software Developer / AI Engineer
Mandrill Tech Sdn Bhd — Kuala Lumpur, Malaysia (Hybrid)
Nov 2020 – July 2025
Leading AI and software engineering initiatives focused on financial technology, biometrics, and
document intelligence. Responsible for designing intelligent systems, automating business
workflows, and deploying ML pipelines in production environments.
Key Contributions:• Financial Advisor Chatbot
◦ Involved in creating the metadata filter for vector search. Created metadata extractor
that extract metadata to be used in metadata filtering during the vector search. The
product search from the vector database improved by 15%
◦ Involved in LLMs and semantic similarity models to automate financial transaction
tagging.
◦ Upserting, Vectorizing the financial product data description into VectorDB.
◦ Involved in migrating the VectorDB from PGvector to Pinecone.
◦ Automate the schema migration for the the chat database.
◦ Involved in deployment into SIT, UAT environment.
• Bank Transaction Classification
◦ Applied LLMs which is BERT and comparing the performance with traditional RNN
model which is LSTM to automate financial transaction tagging. Result showed
BERT model outperforms LSTM just by 2%
◦ Improved categorization accuracy by over 30% and enhanced downstream expense
insights for users.
• Financial Product Recommendation System
◦ Designed hybrid rule-based and score-based recommendation engines for:
▪ Credit card selection (cashback, points).
▪ Refinancing eligibility based on DSR.
▪ Insurance products based on age and financial risk profiles.
◦ Helped improve financial literacy and user conversion rates for clients.
• Financial Mathematical Framework (FMF) API
◦ Developed a financial analytics engine that computes user-level financial health
metrics over time.
◦ Exposed via FastAPI and integrated with banking dashboards.
• NER Pipeline for Textual Financial Data
◦ Generated, labeled, and trained NER models to extract entities such as bank names,
loan types, and monetary values from user chats and scanned documents.
• eKYC System Development
◦ Built a complete KYC verification pipeline using:
▪ Document segmentation.▪ Landmark Detection
▪ Liveliness detection
▪ Face detection and facial matching with landmark analysis.
▪ Blur/glare detection for document quality checks.
▪ OCR for national ID/passport verification (MRZ parsing).
◦ Integrated with client onboarding platforms and passed accuracy benchmarks.
• Mobile OCR App PoC
◦ Built an Android proof-of-concept to scan and read utility meter readings offline
using embedded TensorFlow Lite OCR models.
• Traffic Violation Detection System
◦ Developed deep learning models for red-light detection, vehicle tracking, and license
plate OCR from CCTV feeds.
◦ Reduced false positives using image quality heuristics and temporal consistency.
• Annotation Tooling and Workflow Automation
◦ Created Python scripts for semi-automated annotation using weak supervision,
boosting dataset creation speed.
• GPU Environment Setup
◦ Deployed Dockerized GPU environments for model training.
◦ Reduced training time by 40% through optimization techniques (mixed precision,
data parallelism).
• Schema Migration Tooling
◦ Wrote database migration scripts to manage PostgreSQL schema changes for ML-
integrated applications.
AI Engineer (Freelance)
Hexamatics Servcomm Sdn Bhd — Remote
May 2024 – Nov 2024
Focused on large language model (LLM) applications and enterprise automation for internal HR
and policy management.Key Contributions:
• RAG-Based HR Chatbot
◦ Architected a Retrieval-Augmented Generation (RAG) chatbot using OpenAI,
LangChain, and LlamaIndex.
◦ Enabled HR policy retrieval from internal SOPs and documents using semantic
search.
◦ Built a user-facing FastAPI and Streamlit interface for live queries.
• Query Expansion & Intent Routing
◦ Designed NLP pipelines to identify user intent and route queries to appropriate
response sources.
◦ Improved answer relevance and lowered fallback rate significantly.
• Document Chunking & Indexing
◦ Optimized document chunking strategy for context preservation and accurate
semantic retrieval.
AI Engineer (Freelance)
Popsical Karaoke Sdn Bhd — Remote
May 2021 – Nov 2021
Delivered recommendation algorithms and analytics tools to personalize the user experience on a
karaoke streaming platform.
Key Contributions:
• Music Recommendation Engine
◦ Developed hybrid recommenders combining collaborative filtering, content features,
and popularity metrics.
◦ Tracked user-item interactions and built vector embeddings for similarity scoring.
• Behavioral Data Analysis
◦ Conducted in-depth exploratory data analysis (EDA) to identify user churn, peak
activity times, and genre preferences.
• Behavioral Data Analysis
◦ Conducted in-depth exploratory data analysis (EDA) to identify user churn, peak
activity times, and genre preferences.
• API Development◦ Developed API for the recommendation system using FastAPI.
◦ Run test for inference speed for single and concurrent requests.
Machine Learning Engineer
Blinkware Tech Sdn Bhd — Kuala Lumpur, Malaysia
Sept 2019 – Apr 2020
Worked on computer vision-based safety systems, industrial quality control, and real-time detection
systems for automotive and food industries.
Key Contributions:
• Driver Fatigue Detection
◦ Built MobileNet-SSD and OpenCV-based fatigue detection for real-time blink/eye
closure monitoring.
◦ Increased road safety awareness in collaboration with transport sector clients.
• Dental Image Segmentation
◦ Trained deep learning models for tooth and cavity segmentation using U-Net.
◦ Used in a dental imaging product to assist in diagnostics.
• Packaging Inspection with OCR
◦ Implemented OCR + object detection to verify batch codes and expiry dates on
cereal boxes.
◦ Enabled quality control via visual inspection automation.
• Color Clustering for Quality Control
◦ Used K-Means and color histograms to classify packaging colors and detect
inconsistencies.
Education
Master of Computer Science
Universiti Teknologi MARA (UiTM) — 2018–2020
GPA: 3.60Bachelor of Microelectronic Engineering
Universiti Malaysia Perlis (UniMAP)
Publications
Comparing MobileNet-SSD and YOLO v3 Learning Architecture for Real-Time Driver’s
Fatigue Detection
International Journal of Academic Research in Business and Social Sciences (IJARBSS) — Dec
2021
View Publication
Achievements
• Best Paper Award – CSSR 2020 (Logistics & Transportation Track)
• Faculty Postgraduate Award – Universiti Teknologi MARA (GPA: 3.79)