Egemen TUNCARSLAN

Final-year Computer Engineering student (English track) with a strong applied AI background spanning machine learning, deep learning, NLP, computer vision, and data-centric system development. I have built and deployed production-oriented AI pipelines, including Agentic AI workflows and retrieval-augmented generation (RAG) systems, with an emphasis on reliability, evaluation, and real-world constraints. In parallel, I am pursuing undergraduate thesis research on WiFi CSI–based sensing with latent diffusion models and exploring quantum computing concepts to reduce uncertainty in inference and improve decision confidence. I am applying to San José State University’s MS in Artificial Intelligence to deepen my research skills and contribute to applied AI work at the intersection of robust ML systems and next-generation computing paradigms.

Certificates are available on my LinkedIn profile

1) LatentCSI: Quantum-Enhanced WiFi Sensing for Disaster Response (Undergraduate Thesis)

Jan 2025–Present | Advisor: Prof. Dr. İhsan Yılmaz (Maltepe University)

Platform: NVIDIA Jetson Xavier NX, Intel AX210 (WiFi 6 CSI)

Tech: Python, PyTorch, Stable Diffusion (latent diffusion), FAISS, Qiskit

Implementing and extending a latent diffusion–based CSI→image reconstruction pipeline using a pretrained Stable Diffusion prior (latent-space mapping rather than pixel-space generation).

Evaluating robustness under domain shift (device position, occlusion, temporal drift) and privacy-preserving constraints; observed improved perceptual quality (FID ↓) versus pixel-space baselines in controlled experiments.

Ongoing: exploring hybrid quantum-classical approaches to reduce uncertainty and improve reliability of human-presence verification from CSI for disaster-response scenarios.

Funding: TÜBİTAK 1002-B grant application submitted (₺100K requested; decision pending).

2) AI-Powered Portfolio Optimization System (RoboAdvisor)

Sep 2024–Present | Market: BIST (Turkey)

Tech: Python, PyTorch, XGBoost, TA-Lib, CVXPY

Built a forecasting + signal-generation pipeline (LSTM-based return forecasting; XGBoost for 5-class signal classification) across ~500 equities over ~2,000 trading days.

Integrated ML outputs into a mean-variance optimizer (CVXPY) and implemented walk-forward evaluation with per-asset performance tracking (RMSE, R², directional accuracy).

Applied calibration and stability improvements to align predictions with historical distributions and improve backtesting consistency.

Impact: production-oriented system for risk-adjusted portfolio recommendations in quantitative finance.

3) Enterprise Multi-Modal Financial Assistant Platform (Industry)

Oct–Dec 2024 | Client: Geneks Yazılım A.Ş.

Tech: E5-Large, Mistral-7B (INT4 GGUF), Whisper (Large-v3), FAISS, FastAPI, MongoDB

Built a voice-enabled conversational assistant consolidating 11+ internal financial services/APIs into a single interface for Turkish brokerage workflows.

Implemented a secure pipeline with input sanitization, embedding-based intent routing (85%+ accuracy; internal evaluation), and hybrid RAG with confidence-threshold routing.

Delivered observability (routing distribution, interaction metrics), session-based authentication, and A/B shadow testing; optimized on-device inference via quantization and caching.

Impact: reduced user interaction time by ~60% and enabled voice-first institutional usage.

4) AI-Powered ESG Audit & Compliance Platform (Greendeks Backend) (Optional)

Nov–Dec 2024 | Event: KKB Hackathon 2025

Tech: Flask, PostgreSQL, FAISS, LangChain, multilingual embeddings

Developed an evidence-grounded ESG verification backend using a multi-agent review workflow (advocate/skeptic/judge) to flag potential greenwashing.

Implemented document ingestion, vector search, threshold-based routing, and cross-document verification with auditable evidence trails.

Impact: audit-style outputs with transparent evidence provenance and compliance-ready reporting.

• TOBB Young University Student Council — Founder & Chair, Turkish AI Law Commitee (Türk Yapay Zeka Yasası Çalıştayı)
• TurcoPrepneurs (U.S. Embassy–Affiliated Program) — Entrepreneurship Ecosystem Project; Cryptology Track/Project (Dec 2022 – May 2023)
• Gökçen SİHA — Founding Captain; Head of Computer Vision & AI (Aug 2023 – Dec 2025)
• Türkiye’nin Kıvılcımları — Campus Leader
• BiTekDer – 100s Club — Campus Representative
• TOBB Young University Student Council — Student Member
• Turkish Exporters Assembly (TİM) — Campus Representative; Innovation Ambassador Program Participant
• Maltepe University Sustainability Office & Sustainability Club — Founding team; Vice President (active role during establishment)
• MUN (Model United Nations) Society — Founder & President (1.5 years); 7 awards
• Maltepe University Robotics Club — Founder & President
• VEX Robotics Tournament — Co-Organizer; managed end-to-end operations for a 300-participant event
• TİM Inovatim Innovation Competition — Team Founder
• Maltepe University Mayatto Technology Office — Contributed to the establishment phase
• Maltepe University Science & Technology Club — Audit Board Member
• Maltepe University Literature Club — Audit Board Member
• Maltepe University Dance Club — Member
• Medeniyet University Technopark Networking Speed Event — Moderator

• Robot El Association — Volunteer (Jul 2025 – Present)
• AHBAP — Volunteer (Aug 2024 – Present)
• TEMA Foundation — Volunteer (Dec 2024 – Present)
• WWF — Volunteer (Aug 2025 – Present)
• QTurkey — Volunteer (Sep 2025 – Present)
• Maltepe Municipality & Maltepe University (Partnership Program) — Volunteer Instructor (Nov 2024)
  Delivered digital literacy training for homemakers as part of a university–municipality community education initiative.