Fulbright Principal Candidate, 2026–2027Istanbul, Turkey

Neşenaz
Yalçın

Computer Science & Industrial Engineering, Sabancı University

Research inInformation retrievalEdge AIHumanitarian logistics

Download CVEmail
Information retrieval✦Rank fusion✦Edge AI✦Small language models✦Humanitarian logistics✦Simulation✦

0

publications in 2026

ICMLA 2026 and IEEE SIU 2026

0%

exact-match accuracy

fine-tuned Granite-350M, up from 17.0%

0+

operator commands

synthetic dataset for MQTT tool calls

0

countries

Turkey, Italy, China, Netherlands, South Korea

0

GPA across a double major

Computer Science and Industrial Engineering

0

IELTS overall band

Listening 9.0, Reading 9.0

Publications

Peer-reviewed, 2026

ICMLA 2026 · Accepted

Beyond Skewness: Confidence-Guided Rank Fusion for Text Retrieval

N. Yalçın, E. Teper, M. Keskin, M. Öztürk Umut, Y. H. Şahin

International Conference on Machine Learning and Applications (ICMLA 2026). Accepted.

Extends SkewFuse to heterogeneous text retrieval with per-query score normalization and Top-K Gap, Z-Score, Entropy and Hybrid confidence estimators, evaluated on ArguAna, NFCorpus and SciFact.

Live fusion on this site's index

mean of Top-K gap and entropy

BM255 hits · conf 0.23

w = 0.19

Trigram1 hits · conf 1.00

w = 0.81

  1. 1Data Science Intern
  2. 2Beyond Skewness: Confidence-Guided Rank Fusion for Text Retrieval
  3. 3Data Science Intern
  4. 4Disaster Relief Logistics: Assessing Multinational Brands' Responses
  5. 5Skills
PaperAcceptance
SIU 2026 · IEEE

Edge AI Integration Architecture for the Internet of Things

D. Alkanalka, M. Seçen, N. Yalçın, H. E. Yıldız, E. Çimenoğlu, T. Helvacıoğlu, D. Keküllüoğlu

34th Signal Processing and Communications Applications Conference (SIU 2026), IEEE.

edgecloud
IEEE Xplore

Experience

Research, industry, teaching

ResearchJun 2026 – Aug 2026

University of Twente

Research Intern (Erasmus+ Traineeship)

Enschede, Netherlands

  • Reproduced three published AI/ML papers on 5G/O-RAN network security and handover prediction from their written methodology alone, then diagnosed discrepancies against the authors' released code and models.
  • Synthesized the results into a nine-point Minimum Disclosure Set for ML paper reproducibility, published to the team's documentation site.
Reproducibility5GO-RANNetwork security
TeachingFeb 2026 – Jun 2026

Sabancı University

Learning Assistant, DSA 210 Introduction to Data Science

Istanbul, Turkey

  • Led weekly office hours and one-on-one meetings, guiding students' course projects from data collection and analysis to final presentation.
Data scienceMentoring
IndustryJan 2026 – Jun 2026

Hepsiburada

Data Science Intern

Istanbul, Turkey

  • Developed a query labeling framework using LLMs and applied machine learning to classify search queries as broad, specific or ambiguous.
  • Prototyped agent-based systems to automate experimentation for new search features.
  • Optimized ranking feature weights using NDCG@k and CR@k, targeting search-driven sales.
SearchRankingLLMNDCG
ResearchSep 2025 – Jun 2026

Siemens

Senior Project Student

Istanbul, Turkey

  • Built a resource-efficient, secure Action Orchestration Agent for industrial edge automation using small quantized language models (SLMs) to map natural-language operator commands to protocol-driven MQTT actions locally.
SLMQuantizationMQTTEdge AI
IndustryAug 2025

Bekaert

R&D Intern

Shanghai, China

  • Designed and implemented data extraction and reporting pipelines from .zs2 files for mechanical lab tests, and applied machine learning to classify tensile test outcomes as acceptable or defective.
Data pipelinesClassification
IndustryJun 2025 – Jul 2025

Getirfinans

Data Science Intern

Istanbul, Turkey

  • Automated the extraction and comparison of interest and deposit rates with asynchronous Python and Playwright web scraping pipelines.
PythonPlaywrightWeb scraping
ResearchFeb 2025 – Jun 2026

Sabancı University, SUHOPE

Undergraduate Research Assistant

Istanbul, Turkey

  • Contributed to research projects at the Sustainable Health and Humanitarian Operations Research Lab, working with PhD students and faculty on humanitarian logistics and disaster response.
Humanitarian logisticsDisaster response
ResearchFeb 2024 – Feb 2025

Sabancı University

Research Assistant, Program for Undergraduate Research (PURE)

Istanbul, Turkey

  • Researched disaster relief logistics over three semesters under the supervision of Prof. Raha Akhavan Tabatabaei and Altuğ Tanaltay.
Disaster reliefLogistics

Projects

Models, reports, systems

17.0% → 99.5%

Resource-Efficient AI Action Orchestration Platform for Industrial Edge Devices

  • Benchmarked FunctionGemma-270M, Granite-350M and Gemma 3-1B under zero-shot, few-shot and fine-tuned settings for translating natural-language operator commands into MQTT tool calls, on a synthetic dataset of 3,700+ commands.
  • Developed a method-based intent model that separates action type from device identity, so it scales across sensor configurations.
  • Fine-tuned Granite-350M: exact-match test accuracy rose from 17.0% to 99.5%, with 100% tool-call accuracy and 96.8% correct rejection of out-of-scope commands, matching Gemma 3-1B with roughly one third of the parameters and fitting 4–8 GB VRAM edge hardware.
Fine-tuningSLMTool callingMQTTPyTorchTransformers
Before fine-tuning17.0%
After fine-tuning99.5%
Tool-call accuracy100.0%
Out-of-scope rejected96.8%
Granite-350M
Gemma 3-1B, same accuracy

11 cities, 2 need lists

Improving Coordination in Disaster Relief Donations via News Cycle Analysis

  • After the 2023 Kahramanmaraş earthquake, almost all news outlets relied on two AFAD need lists for eleven affected cities, missing how regional needs shifted over time.
  • The top five communicated needs changed week by week: food and water in the first days, shelter and hygiene products by the fourth week, while over-donation of clothing created supply–demand mismatches.
  • Compared media-reported needs with donor survey data: largely overlapping, with gaps in child-specific and sleeping supplies.
News analysisDonationsSurvey dataReport ↗

Week 1

  1. 1Food & water
  2. 2Winter clothing & heaters

Week 4

  1. 1Shelter
  2. ↑Hygiene products
  3. 5Winter clothing & heaters

60 companies

Corporate Social Responsibility Communication in Disaster Relief

  • Collected and annotated thousands of social media posts from 60 local and international companies after the 2023 Kahramanmaraş earthquake.
  • Classified content as informational, transformational or interactional; emotionally resonant posts with visual media drew the highest engagement.
AnnotationClassificationCSRReport ↗

Disaster Relief Logistics: Assessing Multinational Brands' Responses

  • Analyzed earthquake-related tweets in Python with classification, correlation and regression; disaster-related posts drew higher engagement.
  • Identified sector-specific response patterns, from donations and supply deliveries to free communication and logistics support, and evaluated brands' crisis communication.
PythonRegressionSocial mediaReport ↗

Airport Terminal Simulation & Optimization

  • Built a discrete-event simulation model in Arena to analyze passenger flow and optimize resource allocation.
  • Validated the model with Chi-Square, Kolmogorov–Smirnov and paired t-tests.
ArenaSimulationStatistics

Real-Time Messaging Application

  • Built messaging and friend management with Java, Spring Boot, MongoDB, Docker and JWT authentication, covering registration, sessions and message exchange.
JavaSpring BootMongoDBDockerJWT

Education

Sabancı, Bologna, Sungkyunkwan

Istanbul2022 · Sabancı UniversityBolognaFeb 2025 · Erasmus+ exchangeShanghaiAug 2025 · Bekaert R&DEnschedeJun 2026 · University of TwenteSeoulAug 2026 · Sungkyunkwan University
Oct 2022 – Jun 2026Istanbul, Turkey

Sabancı University

B.S. in Computer Science and Engineering

GPA 3.65 / 4.0

Coursework: Software Engineering, Data Structures, Data Science, Artificial Intelligence, Natural Language Processing, Mobile Application Development.

Oct 2022 – Jan 2027Istanbul, Turkey

Sabancı University

B.S. in Industrial Engineering

GPA 3.65 / 4.0

Coursework: Operations Research I, II & III, Simulation, Statistical Modelling, Decision Economics, Decision Analysis, Digital Transformation.

Aug 2026 – Dec 2026Seoul, South Korea

Sungkyunkwan University

Global Exchange Program

Feb 2025 – Jun 2025Bologna, Italy

Alma Mater Studiorum Università di Bologna

Erasmus+ Exchange Program

Skills

Tools I work with

PythonC++JavaRSQLMongoDBGitDockerMQTTArenaGurobi
scikit-learnTensorFlowPyTorchNLTKTransformersBERTGPTReactSpring BootJWTPostgreSQL

Programming & tools

Python, C++, Java, R, SQL, MongoDB, Git, Docker, MQTT

Machine learning

scikit-learn, TensorFlow, PyTorch, NLTK, Transformers, BERT, GPT

Web

React, Spring Boot, JWT, PostgreSQL

Optimization & simulation

Arena, Gurobi

Languages

Turkish (native), English (C2)

Also built a full-stack e-commerce platform with React, Spring Boot, JWT and PostgreSQL.

Languages: Turkish (native), English (C2)

Service and volunteering

Teaches mathematics to middle-school students from underserved regions with the School Support Association; took part in a civic engagement project supporting refugees, low-income families and women; volunteered with Haytap and local groups caring for injured and displaced animals after the 2023 Kahramanmaraş earthquake.

Awards

Recognition

01 / 04

Fulbright Principal Candidate, 2026–2027

Selected as a principal nominee by the Turkish Fulbright Commission for the Fulbright Foreign Student Program, for graduate study in the United States.

02 / 04

Dean's High Honor List, Sabancı University

Fall 2022–23, Fall 2023–24, Spring 2023–24, Fall 2024–25, Fall 2025–26.

03 / 04

1st Place, Promise for Tomorrow Program

Startup idea competition organized by Sisterslab, an NGO promoting gender equality in STEM, and Hepsiburada.

04 / 04Certificate ↗

IELTS 8.5 / 9.0

Listening 9.0, Reading 9.0, Writing 7.5, Speaking 8.0.

Contact

nesenaz007@gmail.com
Download CVLinkedInGitHub