Nitin Sharma

Consultant

Tübingen, Baden-Württemberg, Germany4 yrs 4 mos experience

Most Likely To SwitchAI ML Practitioner

Key Highlights

Expert in mechanistic interpretability of Large Language Models.
Published researcher in continual learning and LLM interpretability.
Developed innovative frameworks for neuroimaging workflows.

Stackforce AI infers this person is a Healthcare-focused AI Researcher with expertise in neuroimaging and machine learning.

Contact

nitinsharma3150@gmail.com LinkedIn

Skills

Core Skills

Large Language Models (llm)Mechanistic InterpretabilityExplainable AiPython (programming Language)Deep LearningMachine LearningTeaching / InstructionReinforcement LearningTeam LeadershipTeachingNeuroimagingVolunteering

Other Skills

AI EthicsAI SafetyComputational ModelingData AnalysisHidden Markov ModelsLeadershipLeadership DevelopmentLiterature ReviewNatural Language Processing (NLP)Normative modelingProblem SolvingPsychotherapy ApplicationsQ-learningR (Programming Language)Statistical Data Analysis

About

Research Assistant at University of Tübingen focusing on mechanistic interpretability in Large Language Models. Developing steering vectors for knowledge access and contamination-free evaluation frameworks. Published researcher in continual learning and LLM interpretability with expertise in neuroimaging and deep learning.

Experience

Universitätsklinikum tübingen

Research Assistant

Apr 2025 – Present · 11 mos · Tübingen, Baden-Württemberg, Germany

@Mental Health Mapping Lab & Bethge Lab, PI: Dr. Thomas Wolfers and Dr. Çağatay Yıldız
Developing activation engineering techniques for steering vectors in LLMs to access latent knowledge without pre-training.
Created novel framework for domain-specific LLM evaluation addressing contamination issues. Paper: https://arxiv.org/abs/2506.07658
Developing GAMLSS Python package for neuroimaging workflows with parallel processing capabilities.
Applied mechanistic interpretability methods to understand knowledge representation in language models.

Large Language Models (LLM)Mechanistic InterpretabilityExplainable AINormative modelingAI Safety

Neuromatch

Teaching Assistant

Jul 2024 – Jul 2024 · 0 mo · Remote

@Neuromatch Academy, Deep Learning Course
Facilitated daily tutorials and project work for an intensive three-week Deep Learning course, guiding international students.

Teaching / InstructionDeep Learning

Tübingen ai center

Graduate Researcher

Apr 2024 – Mar 2025 · 11 mos · Tübingen, Baden-Württemberg, Germany · On-site

@Bethge Lab, PI: Dr. Çağatay Yıldız
Understanding Knowledge Acquisition Through Mechanistic Interpretability of Pre-training
Interpretable Knowledge Mechanisms via Activation Engineering and Steering Vectors

Mechanistic InterpretabilityNatural Language Processing (NLP)Large Language Models (LLM)

Max planck institute for biological cybernetics

Research Assistant

Nov 2023 – Feb 2024 · 3 mos · Tübingen, Baden-Württemberg, Germany · On-site

@Computational Neuroscience lab, PI: Dr. Sara Ershadmanesh
Investigated meta-cognitive abilities in decision-making tasks, developing Q-learning and HSMM-based models.

Reinforcement LearningHidden Markov ModelsQ-learning

Universitätsklinikum tübingen

2 roles

Research Assistant

Sep 2023 – Nov 2023 · 2 mos · Tübingen, Baden-Württemberg, Germany

@Mental Health Mapping lab, PI: Dr. Thomas Wolfers
Analyzed applications of Large Language Models (LLMs) in psychotherapy, focusing on interpretability and ethical challenges.

Literature ReviewLarge Language Models (LLM)AI Ethics

Research Assistant (HiWi)

Apr 2023 – Mar 2025 · 1 yr 11 mos · Tübingen, Baden-Württemberg, Germany

@Mental Health Mapping lab, PI: Dr. Thomas Wolfers
Contributing to multiple projects applying machine learning and deep learning techniques to neuroimaging and mental health research:
Developing GAMLSS Python package for neuroimaging workflows.
Evaluated explainable AI techniques for large-scale rs-fMRI datasets.
Analyzed LLMs in mental health, focusing on interpretability.
Developed ML models for postoperative delirium prediction. Pre-print: https://doi.org/10.1101/2024.03.13.24303920

Deep LearningExplainable AINormative modelingR (Programming Language)Statistical Data Analysis

Academic reinforcement program - iit roorkee

Teaching Assistant

Jan 2022 – Mar 2022 · 2 mos · Roorkee, Uttarakhand, India

I worked as a Teaching Assistant to assist freshers with the coursework of BT-103 (Computer Systems and Programming). I helped them with programming and theoretical doubts and occasionally took classes to provide a summary of modules.

Python (Programming Language)LeadershipTeaching

Student mentorship program, iit roorkee

Student Mentor

Dec 2021 – Jul 2022 · 7 mos · Roorkee, Uttarakhand, India

The Student Mentorship Program is an initiative under the aegis of the Dean of Students Welfare to help freshmen cope with academic, extra-curricular, and personal problems. I worked with a team of first-year students by taking regular meetings to help them in their academic, personal, and professional careers.

Team LeadershipProblem Solving

Universitätsklinikum jena

Research Fellowship

Jun 2021 – Aug 2021 · 2 mos · Remote

The internship was completed under the WISE scholarship 2021 by 'DAAD: German Academic Exchange Service' for the summer internship in Germany program 2021. It aimed to find biomarkers for Major Depressive Disorder (MDD) under the supervision of Prof. Dr. Martin Walter, Prof. Deepti Bathula, and Dr. Meng Li. During this period, I worked on several neuroimaging tools such as dynamic functional connectivities, Graph theory-based, and RNN-based time series features. We employed various machine learning techniques for feature selection and classification and deep learning frameworks like Artificial neural network (ANN), Long-short term memory (LSTM), Autoencoder, Convolution, etc.

fMRINeuroimagingDeep LearningData Analysis

Nss iit roorkee

Executive

Jul 2019 – Jun 2020 · 11 mos · Roorkee, Uttarakhand, India

I worked as an executive for Pragati Cell NSS for sessions 2019-20 and volunteered at Event Management Cell NSS for 2018-19 in IIT Roorkee. During this period, I help organize 'Blood Donation Camp 2018 and 19, 'Indian Road Safety Campaign,' and 'Food Drive' in collaboration with Khalsa Aid. With the help of team members, we executed the 'Ganga Cleanliness Drive' in the Roorkee canal and areas of Haridwar by collecting plastic waste and set up 'Daan Petika' with Nagar Nigam to collect extra/old clothes from nearby people and distribute them to needy ones.

VolunteeringLeadership DevelopmentTeamwork