imrans.live

Open to software engineering & applied AI rolesChicago, IL

Imran Ahmed.

Software Engineer / Applied AI & Cloud Systems

I build AI applications, backend services, and cloud systems—from agent forensics with Vautyn to shared expenses with SplitAI and personalized travel with ChiGo.

Explore my workView résuméGet in touch
GitHubLinkedIn

Featured project · AI agent forensics

Vautyn

Understand what an AI agent did.

Vautyn connects agent tool calls to AWS CloudTrail events. Investigators can follow the evidence, inspect a supported match, and see where an attribution remains uncertain.

Try the demoEngineering case study

Early product. The public demo uses illustrative data.

Vautyn’s public investigation demo showing an agent tool call, a CloudTrail event, and separate attribution and policy findings.
Live demo screenshot · illustrative recordsEnlarge
PythonFastAPIAWS LambdaCloudTrailS3DynamoDBReactTerraform

Selected work

More featured projects

AI applications with working demos, real users, and award-winning design.

Featured project · Full-stack application

ChiGo

A Chicago itinerary built around you.

A Chicago trip planner that turns traveler preferences into personalized itineraries with real places and interactive activity swaps.

DePaul senior capstone award
Best Use of AI
DePaul senior capstone award
Best Design
Try ChiGoView code
ChiGo’s live trip planner with stay, interests, pace, and budget preferences.
Screenshot from the live siteEnlarge

My contribution

Built Django APIs and PostgreSQL storage to turn a 10-question survey into multi-day itineraries, mapped stops, and one-click swaps.

PythonDjangoPostgreSQLOpenAI APIGoogle PlacesDocker

Featured project · Full-stack application

SplitAI

Shared expenses, simpler settlements.

An expense-sharing app with natural-language entry, receipt capture, and a graph algorithm that simplifies who owes whom.

Active users
15
Average across 1,373 randomized test groups
45% fewer transfers
Try SplitAI
SplitAI’s live website showing an itemized receipt and bill-splitting preview.
Screenshot from the live siteEnlarge

My contribution

Built a Spring Boot API with JWT authentication, a React interface, and Claude-powered expense parsing. Deployed on DigitalOcean with Docker and GitHub Actions.

ReactSpring BootPostgreSQLClaude APIDockerJava

More engineering work

01 / Applied AI research

Cross-Family LLM Distillation

Distilled Claude Sonnet into smaller Qwen2.5 models across five NLP tasks using 12,500 teacher-labeled examples and multi-task LoRA.

My contribution

Built a 4-bit training pipeline and an evaluation harness that blocks model promotion when quality regresses, with versioned prompts and sampling settings.

The 3B model closed 86–99% of the performance gap to the teacher on BANKING77, AG News, and IMDb, including IMDb as an unseen task.

PythonPyTorchHugging FacePEFT / LoRAQwen2.5Claude API
Details in résumé

02 / Computer vision

Handwritten Equation Recognition

A computer vision model that reads five-token handwritten arithmetic equations using a custom CNN with separate prediction heads.

My contribution

Created a synthetic dataset with PIL and MNIST, then trained and evaluated the model in PyTorch.

PythonPyTorchCNNPILNumPy
View code

Building & collaborating

Experience

June 2024 – Aug 2024

Software Development Intern

Black Fox Security Solutions

Aurora, IL

  • Developed Python scripts to automate file-validation checks and flag invalid inputs, supporting repeatable file-upload testing.
  • Performed functional testing of file-upload features to evaluate input handling, identify defects, and verify expected application behavior.
  • Authored technical documentation covering internal development workflows and testing procedures, providing a shared team reference.

Oct 2021 – Sept 2022

IT Support Specialist

Humble Properties

Chicago, IL

  • Resolved 20+ hardware and software issues per week while communicating technical solutions to non-technical users, achieving a 95% satisfaction rate.

A little about me

Engineer. Builder. Problem solver.

Imran Ahmed

Imran Ahmed

Chicago, Illinois

Connect on LinkedIn

I work across Python, Java, and TypeScript to turn ideas into useful software. My projects span AI applications, backend APIs, cloud infrastructure, and computer vision.

With Vautyn, I’m exploring a practical question: how can we connect an AI agent’s actions to evidence an investigator can inspect? I’m also looking for software engineering and applied AI opportunities where I can contribute and keep learning.

Education & certification

DePaul University

MS in Artificial Intelligence

Completed June 2026

BS in Computer Science

GPA: 3.7 · Dean’s List, 6 quarters

Coursework in machine learning, neural networks, algorithms, distributed systems, and databases.

AWS Certified Cloud Practitioner (CLF-C02)

September 2026

Skills in practice

What I work with

Java & React

REST APIs, authentication, expense entry, and a responsive interface in SplitAI.

Explore SplitAI
Languages
Python · Java · TypeScript · C++ · SQL
Backend
FastAPI · Spring Boot · Django · Node.js · Express
Frontend
React · Next.js · HTML · CSS
Cloud
AWS Lambda · EventBridge · SQS · S3 · CloudTrail · STS
Databases
PostgreSQL · DynamoDB · MySQL · MongoDB
Infrastructure & tools
Terraform · Docker · GitHub Actions · CI/CD · Git · Linux
Machine learning
PyTorch · Hugging Face Transformers · PEFT / LoRA · scikit-learn · NumPy · pandas · Claude API · OpenAI API

Open to opportunities

Let’s build something useful.

Hiring for software engineering or applied AI? I’d love to hear about the role and the problems your team is working on.

Email meDownload résumé