Skip to content

Azwaar
Khan

Building software at the intersection of engineering, data, and the world beyond.

Coordinates
2.9457° N · 101.7717° E
Location
Bangi, MY
Temperature
25°C
Time zone
UTC+8

Open to work

Let’s talk →

Full-stack engineering Scientific computing Machine learning

About

I’m a Software Engineer. For close to four years I’ve built the systems businesses actually run on, from enterprise platforms and integrations to the data pipelines behind them. What I’ve come to care about most is the part that has to be right, not the part that looks impressive.

That started in research. I studied Computer Science and Intelligent Systems at Universiti Sains Malaysia (USM), where my work went into drug–target interaction prediction, teaching models to find the rare true cases when almost every example in the data says no. Two published papers came out of it, along with something more useful than either. I learned that software is not only for building products but also for asking questions nobody has answered yet.

Now I’m pushing that work toward scientific computing, through machine learning and geospatial systems, and in time toward space technology, where software stops being the product and becomes the instrument. Imbalance Explorer is the first piece of it, taking my own paper’s results and rebuilding them into something you can interrogate rather than read. This site is where the rest of it will go.

  • B.Sc. Computer ScienceUSM
  • Dean’s List×3
  • Graduate TechnologistMBOT · 2025

Work

01

Imbalance Explorer.

Interactive Analysis of Data Balancing in Drug–Target Interaction Prediction

Imbalance Explorer is an interactive visualisation of how different data-balancing techniques affect drug–target interaction prediction. Built from my undergraduate research, it turns the experimental results into something you can explore directly, comparing model performance, resampling strategies, and the effect of class imbalance.

Type
Research Visualisation
Origin
Undergraduate Research
Year
2026
Status
Live

Stack

  • React
  • TypeScript
  • Vite
  • Tailwind CSS
  • Recharts
Imbalance Explorer landing screen, with the project title set large above a citation block for the underlying paper.
Dashboard showing headline metrics, a class balance breakdown, and a table comparing resampling techniques.
Dashboard showing an F1 score heatmap across models and resampling methods, beside a chart of minority class share.

The research citation sits with the project, not behind it

Research → Data → Visualisation

Experience

Jun 2024 — Present2 yrs 4 mos

Software Engineer.

IFCA MSC BerhadPetaling Jaya, SGR

I build and extend HotelX and EventX, two enterprise platforms running on a microservice architecture, working across the stack in React, Vue, TypeScript, Node.js, GraphQL, and PostgreSQL. Much of the work is integration, wiring payment gateways, IoT access control, and telephony into platforms that have to keep working while guests are checking in. I own several modules end to end, from a client’s requirement through testing and staging to production on Azure DevOps, and I mentor the interns joining the team on React and Node.

  • React
  • Vue.js
  • Node.js
  • Apollo GraphQL
  • TypeScript
  • PostgreSQL
  • Azure DevOps

Dec 2022 — May 20241 yr 6 mos

Software Engineer.

Ashisuto Global TechnologiesBayan Lepas, PNG

I worked in a six-person team alongside the CTO, turning client problems in manufacturing, medical, moulding, and travel into working software. Most of it came down to automation and data capture. I wrote Python tooling that read PDF documents into structured data, and an integration with Monitor ERP that kept records synchronised without anyone retyping them. I also built a travel agency’s mid-office system on the PERN stack, working directly with clients to turn operational requirements into something they would use every day.

  • Python
  • React
  • Express
  • PostgreSQL
  • Monitor ERP

Mar 2021 — Aug 20216 mos

Programmer / Research Assistant.

School of Computer Sciences, USMGelugor, PNG

I researched drug–target interaction prediction, building SVM, Naïve Bayes, and convolutional neural network models on chemical and natural product datasets in Python with scikit-learn, TensorFlow, and Keras. The difficulty was imbalance. Real interactions are rare, which lets a model score well while learning almost nothing, so I evaluated resampling techniques such as SMOTE and ADASYN against that problem. The work grew into the research I later published.

  • Python
  • scikit-learn
  • TensorFlow
  • Keras
  • imbalanced-learn

Research

01

Role
First author
Journal
Molecules
Citation
28(4) · Art. 1663 · MDPI
Year
2023

Comparative Studies on Resampling Techniques in Machine Learning and Deep Learning Models for Drug–Target Interaction Prediction.

Benchmarked resampling techniques across machine learning and deep learning models on imbalanced drug–target data. Random undersampling degraded performance sharply on severely imbalanced classes, which made it unreliable. SVM-SMOTE paired with Random Forest and Gaussian Naïve Bayes held high F1 scores throughout, and a multilayer perceptron scored well even with no resampling at all.

Azwaar Khan Azlim Khan · Nurul Hashimah Ahamed Hassain Malim

DOI 10.3390/molecules28041663 ↗

02

Role
Co-author
Journal
Computing and Artificial Intelligence
Citation
1(1) · Art. 99 · Academic Publishing
Year
2023

LINGO Profiles Fingerprint and Association Rule Mining for Drug–Target Interaction Prediction.

Proposed LINGO Profiles Fingerprint, a molecular descriptor for representing compounds, then used association rule mining to reduce it to its most informative fragments. The reduced fingerprints held the accuracy of the established ECFP4 descriptor while running over 250 times faster, and cleared 80% accuracy on three unseen ChEMBL activity classes.

Muhammad Jaziem Mohamed Javeed · Azwaar Khan Azlim Khan · Nurul Hashimah Ahamed Hassain Malim

DOI 10.59400/cai.v1i1.99 ↗

Skills

Languages

  • TypeScript
  • JavaScript
  • Python
  • SQL

Frontend

  • React
  • Next.js
  • Vue.js
  • Tailwind CSS
  • Vite
  • Recharts
  • HTML
  • CSS

Backend

  • Node.js
  • Express
  • Apollo GraphQL
  • REST APIs
  • Microservices

Databases

  • PostgreSQL
  • TypeORM

Infrastructure

  • Azure
  • Azure DevOpsCI/CD
  • Nginx
  • PM2
  • Vercel
  • Git

Systems integration

  • Payment GatewaysBillplz · PayDollar · Xendit
  • IoT & access controlTTLock · Dahua · Janberg
  • TelephonyPABX
  • ERPMonitor ERP
  • Enterprise servicesSynchrowebKiwire · Kiwire
  • E-invoicing

Data / Research

  • NumPy
  • pandas
  • scikit-learn
  • TensorFlow
  • Keras
  • imbalanced-learn
  • RDKit

Contact