Emre mrkckpln

"Building intelligent systems with a focus on algorithmic efficiency and real-world impact."

🎓 Graduate Software Engineer (June 2026) @ Ostim Technical University

🏢 Rotakur: Unified Fleet Supply and Mobility Platform
Graduation Project — Enterprise B2B SaaS + B2C Marketplace
Rotakur is a production-minded hybrid platform that brings corporate fleet operations, organization-led event logistics, and public mobility booking into one multi-tenant system.
- Unified Operations Core: Designed a single architecture where B2B operations, organization workflows, and B2C marketplace orders run on shared business rules with atomic B2C to B2B transaction bridging.
- Smart Routing and Real-time Mobility: Built OSRM-based route intelligence, waypoint-aware booking, geofence validation, and live WebSocket operational streams for tracking, dispatch, and event-room visibility.
- Finance, ESG, and Compliance Engine: Implemented expected vs realized profitability flows, telemetry-driven distance and CO2 metrics, UBL-TR 2.1 e-invoice generation, and resilient dispatch with retries and DLQ-safe fallbacks.
- Security and Tenant Governance: Enforced strict tenant isolation, role-scoped authorization, anti-BOLA access controls, and separated auth domains for enterprise users, organization users, and marketplace clients.
Stack: Python, FastAPI, SQLAlchemy 2.0, Alembic, Pydantic v2, PostgreSQL, TypeScript, Vue 3, Pinia, WebSockets, Redis, OSRM, Leaflet, MinIO/S3, Google Gemini API, Docker

🛡️ DİDEGAH — Autonomous Perception & Surveillance Engine
Trinity Architecture · Fall Detection · Behavioral Anomaly Scoring · Live WebRTC Dashboard
End-to-end CV pipeline: YOLOv11-OBB (detection) · YOLO-Pose (skeletal analysis) · PatchCore & OBB-LSTM (trajectory & surface anomaly inference). Fully optimized with TensorRT for high-performance edge deployment.
Stack: Python, PyTorch, TensorRT, FastAPI, Next.js, Redis, CUDA

⚖️ LegalLens — AI Contract Analyzer
Neural Network Project
A Chrome Extension detecting risky clauses in legal documents using a custom Bi-LSTM + Attention model and Google Gemini for summarization.
Stack: Python, FastAPI, PyTorch, Chrome Extension

♟️ AI-Powered Chess Analysis Platform
Backend with FastAPI, Frontend with Next.js.
Goal: Developing a specialized AI agent that analyzes matches and provides human-like coaching feedback.

Backend & Architecture: Multi-Tenant B2B/B2C SaaS platforms, Async FastAPI, PostgreSQL, WebSockets, and S3/MinIO cloud storage integration.
Smart Mobility Algorithms: OSRM multi-stop routing, dynamic AI surge pricing, and Haversine-based carpooling matching engines.
Computer Vision & AI: Real-time Object Detection (YOLOv11-OBB), TensorRT Optimization, Temporal Behavioral Analysis (LSTM), and NLP Smart Dispatchers via Google Gemini.
System Optimization: GPU acceleration (CUDA), TensorRT FP16 quantization, and high-concurrency event loops for zero-latency pipelines.

Agentic Engineering: Architecting strict system guidelines and utilizing advanced prompting to orchestrate AI-assisted development across massive codebases.
Strategic Problem Solving: Applying professional-level chess foresight (15 years) to software architecture, anticipating edge cases.
High-Performance Computing: Refining real-time video processing pipelines and optimizing edge-device inference capabilities.

Provide feedback