Penetration testing is largely manual, and most automated tools send target data to third-party cloud services — a non-starter in security-critical environments.

Approach

A LangGraph state machine scans a web or repository target, picks the relevant security tools, and runs them in sequence. An LLM reads the raw output, builds a MITRE ATT&CK–aligned exploit graph, and writes a structured vulnerability report. Everything runs on local Ollama models, streamed live to a Next.js and FastAPI interface.

Challenges

Coordinating non-deterministic model output with deterministic security tooling required strict state management and guardrails, delivered under a zero-data-leakage constraint and a 24-hour deadline.

Outcome

A fully offline pipeline that won first place at BirmingHack 2.0 for Best Use of AI Agents on Arm.

Case 02

SocratEase

Multimodal AI · Speech

Problem

Public-speaking feedback is subjective and rarely available immediately after a session.

Approach

SocratEase analyses a speech video across three channels: computer vision for eye contact and facial expression, audio for fluency scored by an XGBoost model, and transcript analysis for coherence and filler words. The three are combined into a single set of structured feedback.

Challenges

Fusing audio, video and language models into one coherent score within a 24-hour build required careful scoping and a clean separation between capture, analysis and presentation.

Outcome

Built by a team of three, it won first place at BirmingHack 1.0 for Best AI Hack.

Case 03

Camera Calibration

Computer Vision · OpenCV

Problem

Every real lens deviates from the ideal pinhole model, introducing radial and tangential distortion that corrupts measurement and 3D reconstruction.

Approach

The toolkit detects the corners of a 9×6 checkerboard across many photographs and solves for the camera matrix and distortion coefficients, supporting both the pinhole and fisheye models. The recovered parameters undistort still images and video frame-by-frame.

Outcome

Reproducible Jupyter notebooks that calibrate a camera and undistort new footage, built as a prerequisite for computer-vision work at METU's Applied Intelligence Lab.

Case 04

Time-Series Forecasting

Data Science · Internship

Problem

Operational planning needs forward-looking estimates, not just historical reporting.

Approach

An ARIMA model trained on historical operational data at Doğuş Teknoloji, producing projections with confidence intervals so stakeholders can weigh the uncertainty.

Outcome

Historical operational data turned into projections that leadership can plan against.