AI Researcher / NCEPU, Beijing

ALI KHAN AI researcher

I work across computer vision, multimodal learning, and AI systems for research problems that have to survive reality.

Enter the work Start a conversation

AI researcher / NCEPU

Computer vision + AI

Research to systems

Built around evidence

01 / About Ali

Research with a real-world horizon.

01 / Research direction

Computer vision, multimodal learning, and optimization.

A graduate of North China Electric Power University, exploring AI systems for medical imaging, autonomous systems, and real-world forecasting.

02 / Current practice

Research assistant across medical AI and energy systems.

Recent work combines Vision Transformers, 3D-CNNs, Bayesian modelling, graph neural networks, and reproducible experiment pipelines.

03 / Teaching + collaboration

Making technical ideas usable for people.

Python teaching assistant for 160+ undergraduates, with research shaped alongside clinicians, data scientists, and energy engineers.

02 / Research archive

Papers, systems, and questions in motion.

A selected record of published work and manuscripts in medical AI, energy systems, and cyber-physical security.

IEEE / ICOSST 2023

Advancing Epilepsy Disease Classification through Machine Learning and Deep Learning Models Utilizing EEG Data

Co-author / Published Research record

Energies / 2026

ECOGridDR: Blockchain-Enabled Automated Demand Response Optimization and Supplier Management for Residential Buildings

Co-author / Published Research record

Under review / 2026

Brain Tumor Detection Through Deep Learning: A PRISMA-Guided Systematic Review and Quantitative Evidence Map

First author / Manuscript under review Research profile

Under review / 2026

A Secure Privacy-Aware and Safety-Oriented Deep Intrusion Detection Framework for Cyber-Physical Systems

Corresponding author / Manuscript under review Research profile

Under review / 2026

MSAAF-Net: Multi-Scale Adaptive Attention Fusion with Multi-Teacher Distillation for Accurate and Lightweight Brain Tumor Classification

First author / Manuscript under review Research profile

Applied Sciences / 2026

SAED: A Numerical Weather Prediction-Guided Stage-Aware Ensemble Diffusion for Multi-Site Probabilistic Wind and Solar Power Forecasting

Corresponding author / Probabilistic renewable forecasting

A conditional diffusion framework that combines known-future weather guidance with stage-aware denoisers for calibrated multi-site wind and solar forecasts.

Research profile

01 / Toolkit

Skills built for research that ships.

From data preparation to model evaluation, each tool belongs to a clear research or engineering decision.

LLMLarge Language ModelsLanguage reasoning

AGTAgentic AI & AI AgentsTool-using workflows

GENGenerative AICreative systems

NLPNatural Language ProcessingLanguage understanding

CVComputer VisionVisual intelligence

MLOpsExperiment lifecycle

RAGRetrieval-Augmented GenerationGrounded responses

PyTorchDeep learning

TransformersModern architectures

MMMultimodal AIText, image, and audio

PythonResearch computing

GitVersioned work

OpenCVVision tooling

Language AI / generative systems / computer vision / research infrastructure

02 / Education

2023 - 2026MSc, Control Science & Computer EngineeringNorth China Electric Power University / BeijingChinese Government Scholarship (CSC)
2019 - 2023BSc, Mechanical EngineeringNorth China Electric Power University / BeijingBeijing Government Scholarship

03 / PhD direction

Decision intelligence for complex, high-consequence systems.

I aim to pursue doctoral research on AI systems that can reason, retrieve evidence, forecast uncertainty, and act within business and computational workflows. The research question is not simply how to improve a model, but how to design LLM, RAG, agentic, multimodal, and forecasting systems that improve planning, operations, and resource allocation under real constraints.

The intended contribution sits across computer science and business: rigorous evaluation, MLOps, optimisation, human oversight, privacy, and measurable decision quality are treated as first-class parts of the system rather than deployment afterthoughts.

LLM + RAG systems
Agentic decision workflows
Forecasting + optimisation
MLOps + evaluation
Human-in-the-loop governance

Discuss doctoral research

03 / Contact + writing

Follow the research, or start a conversation.

For research collaboration, AI engineering, teaching, and conversations around practical intelligent systems.

m.alikhancb25@gmail.com NCEPU AI Lab

Medium / 01

AI ROI in 2026: From Cool Demos to Real Business Value

Why serious AI work must save time, reduce cost, improve decisions, or create measurable value. Read article

Medium / 02

GPT-5.6 vs Claude Fable 5: Which AI Model Should You Actually Use?

A practical comparison of reliability, reasoning depth, and the workflow fit behind model choice. Read article

Medium / 03

Multimodal AI Explained Simply: The Next Step After Text Chat

How AI can combine text, images, audio, video, charts, and documents in one workflow. Read article

03 / Current builds

Projects in public.

GitHub / AliKhan25X

01tinyllm-lab

Pure Python lab for tokenizers, attention, retrieval, sampling, and evaluation internals.

Python / LLMs 02scholarrag

Local-first RAG research assistant with citations, hybrid retrieval, evaluation, and benchmarks.

RAG / JavaScript 03phd-ai-project-suite

An evolving public suite for PhD-level AI project exploration.

AI / Research

04 / Principle

Models are only as good as their evidence.

Every claim should resolve to a project, experiment, thesis result, or working artifact.

Ali Khan in graduation attire outside North China Electric Power University

Research community gathered in an auditorium

Ali Khan standing at the right of his research community

Ali Khan in graduation attire at the ceremony

01 / 05Graduation / research community

05 / A work in progress

Ali Khan / AI researcher

A record of becoming.

I began in mechanical engineering, where systems were tangible: materials, constraints, forces, and outcomes. Moving into control science and computer engineering changed the scale of those questions, but not their core. I became interested in how a model meets a real decision: how evidence is gathered, uncertainty is handled, and a technical result becomes something people can trust.

My master's work at North China Electric Power University brought those threads together through AI research, teaching, medical imaging, renewable forecasting, and intelligent systems. It taught me that the strongest research is not only accurate. It is legible, reproducible, and grounded in the people and systems it is meant to serve.

I am carrying that perspective forward into doctoral work: building AI systems that are technically rigorous, operationally useful, and accountable to the decisions they influence.

Write to Ali