The CV content is also written out below
education
Massachusetts Institute of Technology, PhD
Bilkent University, BS and MS
research
PhD Research Intern, Layer Health
Self-improving agentic LLM systems for clinical decision guidance with human experts in the loop
PhD Research Intern, Apple
Evaluation benchmarks and methods for multimodal (audio, image, text) reasoning with LLMs
PhD Research Intern, Microsoft
Scalable and efficient representation learning algorithms for the personal adaptive immune system
Graduate Student (PhD), Massachusetts Institute of Technology
LLMs, representation learning, sample-efficient statistical and causal inference, health AI
PhD Fellow, Eric and Wendy Schmidt Center, The Broad Institute of MIT and Harvard
Graduate Student (MS), Bilkent University
Reinforcement learning, Bayesian optimization, multi-armed bandits
Undergrad Senior Project, NanoMagnetics Instruments and Bilkent University
Safe online learning algorithms to tune an Atomic Force Microscope
Undergrad Intern, Robotics and Control Laboratory, University of British Columbia
Deep reinforcement learning for point-of-care ultrasound
Undergrad Intern, DataBoss Security and Analytics
ML-based anomaly detection algorithms for cyber-security
teaching
Instructor, Causal Inference Workshop, Paris AI4Health School
Teaching Assistant, 6.7930/HST.956 ML for Healthcare, MIT
Teaching Assistant, EEE485/585 Statistical Learning and Data Analytics, Bilkent University
Teaching Assistant, EEE212 Microprocessors, Bilkent University
publications
*equal contribution · ID = Ilker Demirel
Conference
-
C12
LLMs can construct powerful representations and streamline sample-efficient supervised learning
[paper]
-
C11
Uncovering bias mechanisms in observational studies
[paper]
-
C10
Prediction-powered causal inferences
[paper]
-
C9
Using LLMs for late multimodal sensor fusion for activity recognition
[paper]
-
C8
Scalable universal T-cell receptor embeddings from adaptive immune repertoires
[paper]
-
C7
Prediction-powered generalization of causal inferences
[paper]
-
C6
Benchmarking observational studies with experimental data under right-censoring
[paper]
-
C5
Association of Eicosanoid Metabolites With Cardiac Structure and Function
[paper]
-
C4
Falsification of internal and external validity in observational studies via conditional moment restrictions
[paper]
-
C3
ESCADA: Efficient safety and context aware dose allocation for precision medicine
[paper]
-
C2
Combinatorial gaussian process bandits with probabilistically triggered arms
[paper]
-
C1
Accuracy limits of distance estimation in visible light systems with RGB LEDs
[paper]
Journal
-
J2
Federated multi-armed bandits under byzantine attacks
[paper]
-
J1
Distance and position estimation in visible light systems with RGB LEDs
[paper]
Preprints
-
P2
Machine learning cross-platform proteomic imputation enables protein quality scoring and replication of epidemiological associations
[paper]
invited talks
MIT IMES Seminar · ShahLab at Stanford · Stripe
LLM rubric representation learning
Paris AI4Health School
Delivered a workshop on causal inference for precision medicine
Harvard Data Science Initiative Causal Seminar
Served as the discussant
MIT LIDS Postdoc Seminar
Prediction-powered causal inference
Uludag University School of Medicine
Personalized type-1 diabetes treatment
honors & awards
PhD Fellowship, Eric and Wendy Schmidt Center at the Broad Institute of MIT and Harvard
5G and Beyond Graduate Fellowship, Vodafone and Bilkent University
Comprehensive Undergraduate Scholarship, Bilkent University
Academic Excellence Scholarship, Turkey Higher Education Student Loan and Housing Board
reviewing
NeurIPS 2025, ICML 2026 — Top Reviewer
JMLR, NeurIPS, ICML, ICLR, AISTATS, AAAI, ML4H, IEEE