Meet the Interns at IBM Research - Israel

Technion | Research Area: Hybrid Cloud

"I’m working on distributed inference serving as part of the llm-d open-source project."

Technion | Research Area: Actionable Interpretability in LLMs

"I’m researching the impact of prompt paraphrasing on model robustness across multiple benchmarks."

Technion | Research Area: AI Technologies

"We are researching the introduction of novel tokens into LLMs with custom training and data generation pipelines that will enable the models to detect and correct their own errors at inference time."

Hebrew University | Research Area: Multi-Modal LLMs - Speech

"I am working on improving speech LLMs using preference optimization and reinforcement learning algorithms."

Technion | Research Area: Agentic AI

"I'm working on an external knowledge injection mechanism for LLM-based multi-agent systems to enforce standard operating procedures, boost policy adherence, and improve trustworthiness."

Technion | Research Area: LLM Inference Optimization

"I’m working on leveraging the internal representations of LLMs to optimize next-token generation in large-scale clusters."

Weizmann Institute | Research Area: Data Classification

"My research focuses on leveraging zero-shot and few-shot learning techniques to enable small models to classify novel types of Personally Identifiable Information (PII) at inference time, even when they were not seen during training."

Technion | Research Area: Quantum Software

"I’ve focused on exploring and developing novel resynthesis strategies for quantum circuits and contributing their implementation to the open-source Qiskit project."

Technion | Research Area: Multimodal RAG

"I’m working on the Granite Vision embedding model. My work centers on investigating positional artifacts in visual embeddings."

Technion | Research Area: AI & NLP

"I make AI agents faster and cheaper without compromising their performance quality."

Technion | Research Area: AI for Code

"I’m exploring the use of LLM-as-a-Judge to evaluate the quality of AI-generated artifacts in code generation and analysis tasks."

Technion | Research Area: AI Platforms

"I’m working on enhancing caching mechanisms to improve efficiency and performance in LLM serving platforms."

Technion | Research Area: Agentic AI

"I focused on building verifiers for tool-calling agents, both for test-time scaling and as reward models for RL – a method that allows better selection of reasoning and tool usage trajectories."

Ben-Gurion University | Research Area: NLP

"I am working on a controllable enterprise simulation platform that generates synthetic organizations using world models and LLMs, enabling dynamic reconfiguration and agent-based interactions."

Ben-Gurion University | Research Area: Retrieval-Augmented Generation (NLP)

"I’m working on learning diverse data representations and integrating them efficiently for enhanced retrieval."

Tel Aviv University | Research Area: Mathematics of Computation

"I'm working on finding a trustworthy benchmarking method for optimization algorithms on CPUs and GPUs to understand how to improve current algorithms and find the best possible classical algorithms."

Technion | Research Area: AI for Code Reviews

"I’m developing an AI agent that labels changes in pull requests to assist with code reviews of large GitHub repositories. I’m also working on creating a benchmark to evaluate the agent, collecting data both manually and automatically."

Technion | Research Area: Hybrid Cloud

"I focused on enhancing the performance of automated Retrieval-Augmented Generation (RAG) across multiple domains."

Technion | Research Area: RAG on LLMs (NLP)

"I'm using RAG and LangGraph with LLMs to automate the extraction of data-related risks from legal documents, providing a clear and concise assessment of these risks and their urgency."

Hebrew University | Research Area: NLP

"I’m working on efficient benchmarking for RAG systems that allows trustworthy evaluation of many RAG configurations even on a limited budget."

Bar Ilan University | Research Area: Hybrid Cloud

"I have been leveraging generative AI to develop an agentic AI connector. This project involves addressing the schema matching problem and implementing the necessary transformations."

Technion | Research Area: AI for Multi-Cloud Networking

"I’m working on leveraging AI and NLP techniques to improve and innovate features within multi-cloud networking tools."

Tel-Aviv University | Research Area: In-Context Learning (ICL)

"My research is about leveraging in-context learning (ICL) as a tool for compressing context within deep learning."

Technion | Research Area: Quality Technologies

"I’m working on developing an evaluation measurement for watsonx code assistant. Specifically, our work is about exploring the use of LLMs for code generation evaluation."

Hebrew University | Research Area: Decision Optimization

“I work on theoretical understanding of transformers in the context of combinatorial optimization.”

Technion | Research Area: AI

"During my internship in the AI department, I designed and developed a general web agent to solve complex tasks using an agentic workflow architecture."

Hebrew University | Research Area: NLP

"I work on developing self-improving and self-criticizing LLMs."

Hebrew University | Research Area: Biomedical foundational models for target discovery

"I set out to explore whether transcriptomic foundation models, like scGPT, truly learn biological relationships through pretraining, particularly actual gene regulatory connections."

Technion | Research Area: Red-Teaming LLM Agents

"Through an approach called red teaming, I’m identifying vulnerabilities in LLM agents and uncovering potential risks to ensure we deliver safer, more reliable systems."

Hebrew University | Research Area: security AI

"I’m developing an optimization algorithm that uses scale manipulation within a graph model to reduce multiplication depth in fully homomorphic encryption (FHE) computations for LLMs that execute on encrypted data."

Tel Aviv University | Research Area: Generative Modelling for Drug Discovery

"I contributed to a computational biology research paper that focuses on developing a multi-domain model for drug discovery by fine-tuning the foundation model to learn interesting downstream tasks."

Ben-Gurion University | Research Area: NLP

"I'm generating data that simulates conversations of an analyst in a risk event chat using LLMs. I'm also evaluating the user's response to the answers received through the chat."

Weizmann Institute | Research Area: Multimodal RAG

"My research focuses on advancing multimodal embedding techniques to improve the performance of Retrieval-Augmented Generation (RAG) systems for document retrieval."

Ben-Gurion University | Research Area: Security

"I’m working on developing a comprehensive dataset of SQL queries to train a model to identify SQL injection vulnerabilities."

Tel Aviv University | Research Area: Generative Modeling for Drug Discovery

"During my internship at IBM, I focused on leveraging graph neural networks for the challenging task of drug discovery. This role gave me the unique opportunity to tackle the challenge of designing novel drugs with state-of-the-art computational methods."

Ben-Gurion University | Research Area: NLP

"I am working on generating synthetic data sets to be served as a training set for Large Language Models. I am generating data sets for various query languages using LLMs and syntax rules. "

Technion | Research Area: Quantum

"I worked on single qubit pulsed quantum gate optimization using Machine Learning methods. It has been awesome to experience how cutting-edge research in quantum computing takes part in IBM's products and applications."

Technion | Research Area: Hybrid Cloud

"During my internship at IBM Research - Israel, my area of research has been the static analysis of software."

Technion | Research Area: Security

"I am an intern in the AI Privacy group. I am working on membership inference attacks for time series models."

Hebrew University | Research Area: Hybrid Cloud

"During my internship in the Hybrid Cloud group, I built employed language models for cardinality estimation of complex queries, which are common in analytics and integration pipelines."

Hebrew University | Research Area: Quantum

"During my internship, I am analyzing the performance of quantum error correction codes in the presence of realistic, physically inspired noise models."

Tel Aviv University | Research Area: Artificial Intelligence

"Working on scene understanding and visual question-answering with graphs and LLMs"

University of Oxford | Research Area: Causal Inference and Explainable Machine Learning

"During my time at IBM, I developed a tree-based model that identifies clusters where natural experiments occur. Our model can characterize overlap and estimate causal effects in an interpretable manner."

Hebrew University | Research Area: Secure Communication, Networking, Applied Cryptography

"I work on secure proxying for servers behind a firewall in Hybrid Cloud"

Hebrew University | Research Area: Deep Learning for Healthcare

"Working on pretraining multi-modal transformers in a self-supervised manner on structured data, MRI images, and genomic data from UK Biobank for improving downstream risk prediction tasks."

Hebrew University | Research Area: Quantum computing

"During my internship, I worked on modelling the experimental noisy dynamics of IBM's superconducting qubits."

Centrale Lille (France) | Research Area: Causal Inference, Machine Learning and Healthcare

"I am estimating how much events are unseen within a medical database so that we can correct causal estimations we could derive from it."

Technion | Research Area: mutation analysis for automatic unit-level differential test generation

"At IBM, I work on introducing an additional tool for unit-level test generation, namely mutation analysis, and measuring its statistical advantage."

Bar Ilan University | Research Area: Multi Document Summarization

"In my work, I try to automatically match between customer's security requirements and cloud standards using innovative NLP interpretability methods. IBM's work environment enables high - level research with great people and experts in their field."

University of Haifa | Research Area: Automatic Recognition of Animal Emotions

""The goal of our project is to build an AI-based application able to interpret flow charts on images and answer questions based on such interpretation. As part of this project we built the first GraphQA dataset."

Technion | Research Area: Speech (IBM) Bias removal and generalization in multi-modal settings (Technion)

"I work on learning data representations that can separate the distinct, informative factors in speech signals such as the speaker information, textual content, etc."

Technion | Research Area: Causal Inference, Machine Learning and Healthcare

"I'm working on evaluation of calibration as a metric for Propensity score correctness."

University of Haifa | Research Area: Computer Vision, Machine / Deep Learning, and Compression of Big Data

"I am working on disease progression prediction and simulation in Mammography images."

Tel Aviv University | Research Area: NLP

"I'm working on end-to-end dialog systems."

University of Haifa | Research Area: Machine learning

"I use Machine Learning to identify possible future melanoma and kidney cancer patients"

Technion | Research Area: Identity centric networking and edge systems

"I work on Provider-side network-aware optimization in the 5G environment."

Technion | Research Area: Quantum Information

"I work on Quantum Computing Summer Intern, focusing on quantum state tomography."

Technion | Research Area: AI and Cybersecurity

"I'm working on analytics over encrypted data using fully homomorphic encryption."

Technion | Research Area: Distributed systems and blockchains

"I work on Ceph, an open-source distributed storage platform."

Technion | Research Area: Decision Optimization

"I enjoy working at IBM very much. The people are pleasant and helpful, and the developing possibilities seem interesting and plentiful."

Technion | Research Area: Deep Learning & Computer Vision for Healthcare

"I am working with the Deep Learning & Computer Vision team to win a 3d kidney disease segmentation competition."

Technion | Research Area: Networking ,Cloud, Algorithms

"I work on Function Chain Transport Optimization in Hybrid Cloud."

Technion | Research Area: Neural Networks, Bayesian Deep Learning, Machine Learning

"I am an intern in the AI in Healthcare group. I am working on the topic of multimodal learning using Supervised Contrastive Learning."

University of Haifa | Research Area: NLP

"I work on a project aiming to detect the state of a chatbot-client dialogue at an early stage of the conversation, using classic machine learning methods ad state-of-the-art methods as well."