Afiq CM here. あざっすあざっす[azassu]slangCasual contraction of arigatou gozaimasu. Slang for "thanks," used here as a casual appreciation for checking out my website.
By day, I work with data: refining feature stores, building models, crafting pipeline workflows, and scaling data insight solutions.
By night, I write about the parts of engineering they don't tell you to your face: being a first-generation professional, managing upwards, having agency, navigating local corporate structures, and the invisible dynamics of a career.
About
2012: Performing in Kangar with the varsity brass band, not knowing how my future would unveil itself.
2018: Registered for MPhil and welcomed my daughter, starting the early transition away from electronics engineering.
2021: Saying goodbye to one of the best teams I've worked with. A reminder that in any job, stint, or gig, it is the people you build with that matter most.
Hello there! I'm Afiq. Afiq CM. I've been an early member of founding machine learning & data teams across various companies. Currently I'm in fintech working on tabular ML, but my experience spans propensity models at CIMB, recommendation systems at Popsical, and ML and computer vision systems at Ørsted, BAT, and Moovita, as well as freelancing with a Japanese startup. I've always been keen to bridge the gap between business & data science, as well as making a solution workable & scalable.
I've always loved AI/ML research and maths to an extent, as well as the impact of AI on our daily lives. More recently, I've been exploring AI agents and NLP. Outside of tech, I've spent time ghostwriting travel articles for online platforms. In another life, I probably would have ended up as a linguist, a dancer (personally I think it's a stretch myself, but hey - one can dream!), or a music creator.
Current Interests & Focus
60% Technical
- Tabular ML & deployment
- Machine learning frameworks (Scikit-Learn, PyTorch, TensorFlow)
- NLP & autonomous AI agents
Thoughts
A digital garden mapping the intersection of data, engineering, and career realities.
Work
Selected systems, analytics pipelines, and engineering projects.
-
Cloud Engineering 01
Webhook Ingestion for Demand-Control Alerts
Built a Cloud Run webhook API with BigQuery MERGE upserts on a composite key for idempotent ingestion of real-time demand-control alerts, externalizing the schema to config for safe evolution. Read about the ingestion design →
-
Propensity Modeling 02
Consumer Loan Propensity Model & On-Prem Feature Store
Built ensemble propensity models and pruned local feature stores down to 20 key features to target personal loan leads under strict on-premise memory constraints. Read the project post-mortem →
-
Time-Series Forecasting 03
Sinusoidal Features for Electricity Power Trading
Implemented cyclical trigonometric feature engineering to resolve boundary errors in regional load forecasting, helping maintain stable forecast accuracy for energy trading teams. Read the forecasting analysis →
-
Predictive Modeling 04
Pricing and Volume Forecasting Models
Worked on volume forecasting and pricing models using constrained regression, integrating Shopify demographics and weather inputs with a price snake dashboard for global teams. Read the pricing model exploration →
-
Generative AI 05
Building Occupancy Simulation using Generative Adversarial Networks
Used Generative Adversarial Networks (GANs) during my MPhil research to simulate time-series room occupancy patterns for smart building energy systems. Read the research notes →
-
Autonomous Vehicles 06
LiDAR-Based AV Object Detection Pipeline
Worked with sparse LiDAR point clouds using SqueezeSeg and PointPillars, building pipelines to clean sensor noise and detect 3D objects for autonomous vehicles. Read about working with point clouds →
Currently refreshing my deep-dive case studies. In the meantime, check out my latest notes on the blog.