ML / AI+SOFTWARE ENGINEERINGUIUC · CHAMPAIGN

I build intelligentsystems—from model to interface.

I’m Amaan Khan, a UIUC master’s student and engineer working across computer vision, machine learning, data-intensive products, and reliable backend systems.

AKMODEL → SYSTEM → PRODUCT
Node 1Node 2Node 3Node 4Node 5Node 6Node 7
DATA / AMBIGUITY
DECISION / CLARITY
STACKPY · TS · SQL
1,000+student questions resolved
31,158vehicle detections evaluated
1,000+live markets processed
40%workflow efficiency improvement
00 / POINT OF VIEW

I like the hard middle: turning research ideas into systems people can actually use, measure, and trust.

01 / SELECTED WORK

Built, measured, and made real.

Projects spanning computer vision research, real-time market analytics, and LLM-assisted information retrieval.

01COMPUTER VISION · RESEARCH31,158 detections
Geometry-aware traffic risk compared with the image-plane baseline

Geometry-aware traffic risk estimation.

A four-stage monocular-video pipeline that detects vehicles, maintains trajectories, projects motion into bird’s-eye view, and scores pairwise collision risk.

31,158 detections1,090 frames0.114 pair Jaccard
PythonYOLOv8ByteTrackOpenCVHomography
Read the case study
02FULL-STACK · MARKET DATA1,000+ live markets
MARKETS / REAL TIME

Kalshi prediction market analytics.

A responsive analytics platform for live prediction markets, pairing a Next.js interface and interactive charts with a Python/FastAPI data layer.

1,000+ live marketsSub-second API10+ interactive charts
Next.jsReactTypeScriptPythonFastAPI
03NLP · INFORMATION RETRIEVAL50+ papers per query
RETRIEVE / RESCORE

MiniAsta paper reranking.

A research-paper discovery pipeline that retrieves candidates, applies user constraints, and uses an LLM scorer to rerank for topic fit, method alignment, and practical usefulness.

50+ papers per queryConstraint-aware scoringWeighted reranking
PythonSemantic ScholarLLMsNLPReranking
02 / EXPERIENCE

Teaching, researching, shipping.

A through-line from backend software to applied ML research and helping other engineers learn.

01AUG 2023 — PRESENT

Data Science Course Assistant

STAT 107 & CS 411 · UIUC

Supporting students across data science, statistical modeling, databases, and machine learning while helping refine the learning experience.

Champaign, IL
  • Resolved 1,000+ student support tickets
  • Built and refined Jupyter labs and microprojects
  • Collaborated on curriculum design and assessments
02JAN 2024 — MAY 2024

Machine Learning Researcher

University of Illinois Urbana-Champaign

Led a web-based ML application that used financial data to predict loan approvals with an explicit focus on fairness and accuracy.

Champaign, IL
  • Built the ETL pipeline and PyTorch models
  • Evaluated fairness and accuracy metrics
  • Shipped an interactive JavaScript and SQL interface
03AUG 2021 — AUG 2023

Software Engineer

FOUNDERS

Worked across a public-facing web experience, workflow automation, and backend services for an entrepreneurship community.

Champaign, IL
  • Built Python and Flask backend services
  • Automated event registration workflows
  • Improved operational efficiency by 40%
Programming

Python · TypeScript · C++ · Java · SQL · R · Bash

ML / AI

PyTorch · TensorFlow · Scikit-learn · Transformers · XGBoost · OpenCV

Systems

FastAPI · Flask · Docker · Kubernetes · Airflow · Apache Spark

Cloud

AWS · SageMaker · Google Cloud · Vertex AI · BigQuery

MCS
’27
03 / ABOUT

An engineer who likes both the whiteboard and the profiler.

I’m pursuing a Master of Computer Science at UIUC after completing a B.S. in Information Sciences with a Computer Science minor. I’m happiest when the work asks for both mathematical reasoning and product judgment.

SELECTED COURSEWORK

Artificial Intelligence · Deep Learning · Computer Vision · Machine Learning · Database Systems · Data Mining · Software Engineering

04 / CONTACT

Let’s build something worth caring about.

I’m interested in ambitious ML/AI and software engineering work with thoughtful people. Tell me what you’re solving.

Usually replies within a few days