Honors and Awards

  • 2025.12 Selected as Mentor for the AAAI Undergraduate Consortium 2026, returning after being a Scholar in the 2023 cohort.
  • 2023.02 One of the 11 Undergraduates Selected as an AAAI Undergraduate Consortium Scholar 2023. Included a Travel Grant of $2000 to present at AAAI-23 at Washington DC, USA.
  • 2023.01 Received MAHE Undergraduate Research Grant Worth 10K INR for Project : Explainable & Trustworthy Skin Lesion Classification under Dr. Abhilash K. Pai, Dept. of DSCA, Manipal Institute of Technology, MAHE.
  • 2022.06 Top 10 Team out of 1000+ submissions in Bajaj Finserv HackRx3.0 Hackathon.

Education

  • 2020.10 - 2024.07, Bachelors of Technology in Data Science & Engineering (B.Tech), Manipal Academy of Higher Education, Manipal, Karnataka, India.
    • CGPA: 8.31/10
    • Final Year Project: Pinpointing Oil and Gas Wells from Satellite Imagery using Machine Learning. Internal Guide: Dr. Abhilash K. Pai (Dept. of DSCA, MIT); External Guide: Dr. David Rolnick (Mila, McGill University, UdeM)
Final year project defense slide: Pinpointing Oil and Gas Wells from Satellite Imagery using Machine Learning, with Mila and Manipal Institute of Technology branding
Final Year Project Defense, Manipal Institute of Technology

Academic Service

  • Mentor: AAAI Undergraduate Consortium 2026 (Scholar 2023 → Mentor 2026)

Conference Reviewing & Program Committee

  • Main Conference Reviewer: AAAI 2027 (Program Committee), NeurIPS 2026, ACM AIES 2026, CVPR (2025-26), ECCV (2024-26), ICCV (2025), WACV (2026-27), BMVC 2026, IJCNN (2025)
  • Workshop & Track Reviewer (2026):
    • Mechanistic Interpretability Workshop (ICML 2026)
    • AI for Good Workshop (ICML 2026)
    • TAIGR Workshop (ICML 2026)
    • Foundation Models for Structured Data (FMSD) Workshop (ICML 2026)
    • FinDS Workshop @ ACM SIGMOD 2026
    • Advances in Financial AI Workshop (ICLR 2026)
    • BlackboxNLP Workshop (EMNLP 2026)
    • NLP4PI Workshop (EMNLP 2026)
    • System Demonstrations Track (EMNLP 2026)
    • Actionable Interpretability Workshop (COLM 2026)
    • FAIMI-BRIDGE-EPIMI Workshop (MICCAI 2026)
  • Workshop Reviewer (2021–25):
    • Actionable Interpretability Workshop (ICML 2025)
    • RegML Workshop (NeurIPS 2025)
    • NLP for Positive Impact Workshop (EMNLP 2024)
    • Bayesian Decision-making and Uncertainty Workshop (NeurIPS 2024)
    • Frontiers in Probabilistic Inference (ICLR 2025)
    • SyntheticData4ML Workshop (NeurIPS 2022-23)
    • FAIMI Workshop (MICCAI 2024)
    • Domain Adaptation and Representation Transfer Workshop (MICCAI 2023)
    • Topological, Algebraic, and Geometric P.R.A. Workshop (CVPR 2023)

Technical Skills

Interpretability

Mechanistic interpretability circuit analysis & attribution TransformerLens Captum SHAP LIME Integrated Gradients LRP DL-Backtrace

Post-Training & Alignment

RLHF DPO Constitutional AI PEFT (LoRA, QLoRA, DoRA) TRL (SFTTrainer, DPOTrainer) Axolotl Unsloth LlamaFactory HF Accelerate

LLM Systems & Efficiency

Flash Attention 2 MoE architectures quantization (bitsandbytes, AutoGPTQ) pruning ONNX TensorRT lm-evaluation-harness

Deep Learning Systems

CUDA programming Triton kernels distributed training (DeepSpeed, FSDP) mixed precision GPUs: H200 / H100 / RTX 6000 Pro (multi-GPU)

Tabular & Classical ML

XGBoost LightGBM CatBoost TabPFN Optuna Ray Tune

AI & Coding Agents

LangChain LangGraph Claude Code Codex Cursor Windsurf

MLOps, Serving & Infra

Weights & Biases MLflow Ray Docker Slurm Git Linux HPC HF Hub/Datasets FAISS AWS/GCP vLLM FastAPI Flask Gradio Streamlit Ollama

Languages & Core Libraries

Python C++ SQL Java C LaTeX PyTorch TensorFlow Keras Scikit-Learn NumPy Pandas Seaborn Matplotlib OpenCV PIL Geopandas Shapely NLTK SpaCy

Certifications

  • Deep Learning Specialization - DeepLearning.ai
  • 6th Summer School on AI - CVIT IIITH

Collaborators & Mentees

Full list of managers, mentors, mentees, and collaborators by institution

Numbers in parentheses indicate co-authored works together.

Vinay Kumar Sankarapu (29) MANAGER LEXSI LABS / ARYAXAI ALIGNMENT LABS
  • AlignTune: Modular Toolkit for Post-Training Alignment of Large Language Models
  • C-ΔΘ: Circuit-Restricted Weight Arithmetic for Selective Refusal
  • ALIGNBEAM: Inference-Time Alignment Transfer via Cross-Vocabulary Logit Mixing
  • Forgetting That Sticks: Quantization-Permanent Unlearning via Circuit Attribution
  • Position: Behavioural Assurance Cannot Verify the Safety Claims Governance Now Demands
  • What Do Compliance Detectors Read? An Audit of Activation Probes and Guard Models
  • SafeTune: A Unified Faithful Library for Auditing and Repairing Safety Drift in Fine-Tuned LLMs
  • Drift Then Repair: A Controlled Cross-Paradigm Audit of Safety in Fine-Tuned LLMs
  • Self-Calibrating Weight-Arithmetic Safety-Drift Repair
  • CuratorKIT: Data Curation and Synthetic Data Generation for LLM Post-Training
  • Provenance-Grounded Gating and Adaptive Recovery in Synthetic Post-Training Data Curation
  • Document-as-Function: Verifiable Generation of Long-Form Synthetic Documents
  • Interpretability as Alignment: Making Internal Understanding a Design Principle
  • Interpretability-aware pruning for efficient medical image analysis
  • Bridging the gap in XAI-why reliable metrics matter for explainability and compliance
  • DLBacktrace: A Model Agnostic Explainability for any Deep Learning Models
  • xai_evals: A Framework for Evaluating Post-Hoc Local Explanation Methods
  • CircuitKIT: Circuit Discovery, Evaluation, and Application Toolkit for Mechanistic Interpretability
  • Faithfulness to Refusal: A Causal Audit of Neuron Selectors in LLMs
  • Faithfulness Is Not Actionability: Component Heterogeneity in Discovered Circuits
  • DLBacktrace v2: Extending Model-Agnostic Interpretability for LLMs and MoEs with CUDA Acceleration
  • Orion-MSP: Multi-Scale Sparse Attention for Tabular In-Context Learning
  • Orion-BiX: Bi-Axial Attention for Tabular In-Context Learning
  • Exploring Fine-Tuning for Tabular Foundation Models
  • TabTune: A Unified Library for Inference and Fine-Tuning Tabular Foundation Models (Demo)
  • Pocket Foundation Models: Distilling TFMs into CPU-Ready Gradient-Boosted Trees
  • Ensembling Tabular Foundation Models: A Diversity Ceiling and a Calibration Trap
  • Distilling Tabular Foundation Models for Structured Health Data
  • Data Presentation over Architecture: Resampling Strategies for Credit Risk Prediction with Tabular Foundation Models
Aditya Kasliwal (10) MENTEE MARS ROVER MANIPAL / RESEARCH SOCIETY MANIPAL / LEXSI LABS
  • C-ΔΘ: Circuit-Restricted Weight Arithmetic for Selective Refusal
  • CircuitKIT: Circuit Discovery, Evaluation, and Application Toolkit for Mechanistic Interpretability
  • Faithfulness Is Not Actionability: Component Heterogeneity in Discovered Circuits
  • Performance Evaluation of Deep Segmentation Models for Contrails Detection
  • Sailing Through Spectra: Unveiling the Potential of Multi-Spectral Information in Marine Debris Segmentation
  • ReFuSeg: Regularized Multi-Modal Fusion for Precise Brain Tumour Segmentation
  • AgriLLM: Harnessing transformers for farmer queries.
  • CoReFusion: Contrastive Regularized Fusion for Guided Thermal Super-Resolution
  • LaMAR: Laplacian Pyramid for Multimodal Adaptive Super Resolution (Student Abstract)
  • Laplacian reconstructive network for guided thermal super-resolution
Mohamed Bouadi (8) LEXSI LABS, PARIS
  • Orion-MSP: Multi-Scale Sparse Attention for Tabular In-Context Learning
  • Orion-BiX: Bi-Axial Attention for Tabular In-Context Learning
  • Exploring Fine-Tuning for Tabular Foundation Models
  • TabTune: A Unified Library for Inference and Fine-Tuning Tabular Foundation Models (Demo)
  • Pocket Foundation Models: Distilling TFMs into CPU-Ready Gradient-Boosted Trees
  • Ensembling Tabular Foundation Models: A Diversity Ceiling and a Calibration Trap
  • Distilling Tabular Foundation Models for Structured Health Data
  • Data Presentation over Architecture: Resampling Strategies for Credit Risk Prediction with Tabular Foundation Models
Aditya Tanna (8) MENTEE LEXSI LABS
  • Orion-MSP: Multi-Scale Sparse Attention for Tabular In-Context Learning
  • Orion-BiX: Bi-Axial Attention for Tabular In-Context Learning
  • Exploring Fine-Tuning for Tabular Foundation Models
  • TabTune: A Unified Library for Inference and Fine-Tuning Tabular Foundation Models (Demo)
  • Pocket Foundation Models: Distilling TFMs into CPU-Ready Gradient-Boosted Trees
  • Ensembling Tabular Foundation Models: A Diversity Ceiling and a Calibration Trap
  • Distilling Tabular Foundation Models for Structured Health Data
  • Data Presentation over Architecture: Resampling Strategies for Credit Risk Prediction with Tabular Foundation Models
Soham Bhattacharjee (4) MENTEE LEXSI LABS
  • AlignTune: Modular Toolkit for Post-Training Alignment of Large Language Models
  • CuratorKIT: Data Curation and Synthetic Data Generation for LLM Post-Training
  • Provenance-Grounded Gating and Adaptive Recovery in Synthetic Post-Training Data Curation
  • Document-as-Function: Verifiable Generation of Long-Form Synthetic Documents
Saisab Sadhu (4) MENTEE LEXSI LABS
  • Forgetting That Sticks: Quantization-Permanent Unlearning via Circuit Attribution
  • What Do Compliance Detectors Read? An Audit of Activation Probes and Guard Models
  • SafeTune: A Unified Faithful Library for Auditing and Repairing Safety Drift in Fine-Tuned LLMs
  • Drift Then Repair: A Controlled Cross-Paradigm Audit of Safety in Fine-Tuned LLMs
Neeraj Kumar Singh (4) ARYAXAI ALIGNMENT LABS
  • Interpretability-aware pruning for efficient medical image analysis
  • DLBacktrace: A Model Agnostic Explainability for any Deep Learning Models
  • xai_evals: A Framework for Evaluating Post-Hoc Local Explanation Methods
  • DLBacktrace v2: Extending Model-Agnostic Interpretability for LLMs and MoEs with CUDA Acceleration
Chintan Chitroda (4) MANAGER ARYAXAI ALIGNMENT LABS
  • Interpretability-aware pruning for efficient medical image analysis
  • DLBacktrace: A Model Agnostic Explainability for any Deep Learning Models
  • xai_evals: A Framework for Evaluating Post-Hoc Local Explanation Methods
  • DLBacktrace v2: Extending Model-Agnostic Interpretability for LLMs and MoEs with CUDA Acceleration
Nassim Bouarour (4) LEXSI LABS, PARIS
  • Pocket Foundation Models: Distilling TFMs into CPU-Ready Gradient-Boosted Trees
  • Ensembling Tabular Foundation Models: A Diversity Ceiling and a Calibration Trap
  • Distilling Tabular Foundation Models for Structured Health Data
  • Data Presentation over Architecture: Resampling Strategies for Credit Risk Prediction with Tabular Foundation Models
Akshat Bhandari (4) MARS ROVER MANIPAL
  • Performance Evaluation of Deep Segmentation Models for Contrails Detection
  • UATTA-ENS: Uncertainty Aware Test Time Augmented Ensemble for PIRC Diabetic Retinopathy Detection
  • Evaluating Predictive Uncertainty and Robustness to Distributional Shift Using Real World Data
  • Analyzing Effects of Fake Training Data on the Performance of Deep Learning Systems
Sriya Rallabandi (4) MARS ROVER MANIPAL
  • Performance Evaluation of Deep Segmentation Models for Contrails Detection
  • SSS at SemEval-2023 Task 10: Explainable Detection of Online Sexism using Majority Voted Fine-Tuned Transformers
  • CoReFusion: Contrastive Regularized Fusion for Guided Thermal Super-Resolution
  • LaMAR: Laplacian Pyramid for Multimodal Adaptive Super Resolution (Student Abstract)
Utsav Avaiya (3) MENTEE LEXSI LABS
  • AlignTune: Modular Toolkit for Post-Training Alignment of Large Language Models
  • Faithfulness to Refusal: A Causal Audit of Neuron Selectors in LLMs
  • TabTune: A Unified Library for Inference and Fine-Tuning Tabular Foundation Models (Demo)
Karun Sharma (3) MENTEE LEXSI LABS
  • CuratorKIT: Data Curation and Synthetic Data Generation for LLM Post-Training
  • Provenance-Grounded Gating and Adaptive Recovery in Synthetic Post-Training Data Curation
  • Document-as-Function: Verifiable Generation of Long-Form Synthetic Documents
Sanchit Singhal (3) MARS ROVER MANIPAL
  • Performance Evaluation of Deep Segmentation Models for Contrails Detection
  • SSS at SemEval-2023 Task 10: Explainable Detection of Online Sexism using Majority Voted Fine-Tuned Transformers
  • CoReFusion: Contrastive Regularized Fusion for Guided Thermal Super-Resolution
Chirag Chawla (2) MENTEE LEXSI LABS
  • AlignTune: Modular Toolkit for Post-Training Alignment of Large Language Models
  • ALIGNBEAM: Inference-Time Alignment Transfer via Cross-Vocabulary Logit Mixing
Aadit Sengupta (2) MENTEE LEXSI LABS
  • What Do Compliance Detectors Read? An Audit of Activation Probes and Guard Models
  • Interpretability as Alignment: Making Internal Understanding a Design Principle
Anshul Kaushal (2) MENTEE LEXSI LABS
  • SafeTune: A Unified Faithful Library for Auditing and Repairing Safety Drift in Fine-Tuned LLMs
  • Drift Then Repair: A Controlled Cross-Paradigm Audit of Safety in Fine-Tuned LLMs
Yashwardhan Rathore (2) ARYAXAI ALIGNMENT LABS
  • DLBacktrace: A Model Agnostic Explainability for any Deep Learning Models
  • xai_evals: A Framework for Evaluating Post-Hoc Local Explanation Methods
Hem Gosalia (2) MENTEE LEXSI LABS
  • CircuitKIT: Circuit Discovery, Evaluation, and Application Toolkit for Mechanistic Interpretability
  • Faithfulness Is Not Actionability: Component Heterogeneity in Discovered Circuits
Adil Khan (2) MARS ROVER MANIPAL
  • ReFuSeg: Regularized Multi-Modal Fusion for Precise Brain Tumour Segmentation
  • UATTA-ENS: Uncertainty Aware Test Time Augmented Ensemble for PIRC Diabetic Retinopathy Detection
Kumud Lakara (2) MARS ROVER MANIPAL
  • Evaluating Predictive Uncertainty and Robustness to Distributional Shift Using Real World Data
  • Analyzing Effects of Fake Training Data on the Performance of Deep Learning Systems
Ujjwal Verma (2) MENTOR MIT MAHE
  • Evaluating Predictive Uncertainty and Robustness to Distributional Shift Using Real World Data
  • Laplacian reconstructive network for guided thermal super-resolution
Aryan Kamani (2) MARS ROVER MANIPAL
  • LaMAR: Laplacian Pyramid for Multimodal Adaptive Super Resolution (Student Abstract)
  • Laplacian reconstructive network for guided thermal super-resolution
Ishaan Gakhar (2) MARS ROVER MANIPAL
  • LaMAR: Laplacian Pyramid for Multimodal Adaptive Super Resolution (Student Abstract)
  • Laplacian reconstructive network for guided thermal super-resolution
R E Zera Marveen Lyngkhoi (1) MENTEE LEXSI LABS
  • AlignTune: Modular Toolkit for Post-Training Alignment of Large Language Models
Mykola Khandoga (1) LEXSI LABS, PARIS
  • AlignTune: Modular Toolkit for Post-Training Alignment of Large Language Models
Rui Yuan (1) LEXSI LABS, PARIS
  • AlignTune: Modular Toolkit for Post-Training Alignment of Large Language Models
Nikita Malik (1) MENTEE LEXSI LABS
  • Interpretability-aware pruning for efficient medical image analysis
Ananth Eswar (1) MENTEE LEXSI LABS
  • Faithfulness to Refusal: A Causal Audit of Neuron Selectors in LLMs
Omkar Kakade (1) MENTEE LEXSI LABS
  • DLBacktrace v2: Extending Model-Agnostic Interpretability for LLMs and MoEs with CUDA Acceleration
Yash Jignesh Desai (1) LEXSI LABS, PARIS
  • Ensembling Tabular Foundation Models: A Diversity Ceiling and a Calibration Trap
Mitul Solanki (1) MENTEE LEXSI LABS
  • Data Presentation over Architecture: Resampling Strategies for Credit Risk Prediction with Tabular Foundation Models
Michelle Lin (1) MCGILL UNIVERSITY
  • Alberta Wells Dataset: Pinpointing Oil and Gas Wells from Satellite Imagery
Brefo Dwamena Yaw (1) MCGILL UNIVERSITY
  • Alberta Wells Dataset: Pinpointing Oil and Gas Wells from Satellite Imagery
Jade Boutot (1) MCGILL UNIVERSITY
  • Alberta Wells Dataset: Pinpointing Oil and Gas Wells from Satellite Imagery
Mary Kang (1) MCGILL UNIVERSITY
  • Alberta Wells Dataset: Pinpointing Oil and Gas Wells from Satellite Imagery
David Rolnick (1) MENTOR MILA / MCGILL
  • Alberta Wells Dataset: Pinpointing Oil and Gas Wells from Satellite Imagery
Dyutit Mohanty (1) RESEARCH SOCIETY MANIPAL
  • Sailing Through Spectra: Unveiling the Potential of Multi-Spectral Information in Marine Debris Segmentation
Bharath Udapa (1) RESEARCH SOCIETY MANIPAL
  • Sailing Through Spectra: Unveiling the Potential of Multi-Spectral Information in Marine Debris Segmentation
Siddhant Bharadwaj (1) MIT MANIPAL / IISC BANGALORE
  • Obscure to Observe: A Lesion-Aware MAE for Glaucoma Detection from Retinal Context
Chandra Sekhar Seelamantula (1) MIT MANIPAL / IISC BANGALORE
  • Obscure to Observe: A Lesion-Aware MAE for Glaucoma Detection from Retinal Context
Abhilash K Pai (1) MENTOR MIT MAHE
  • Does the Fairness of Your Pre-Training Hold Up? Examining the Influence of Pre-Training Techniques on Skin Tone Bias in Skin Lesion Classification
Sankarshanaa Sagaram (1) MARS ROVER MANIPAL
  • ReFuSeg: Regularized Multi-Modal Fusion for Precise Brain Tumour Segmentation
Laven Srivastava (1) MARS ROVER MANIPAL
  • ReFuSeg: Regularized Multi-Modal Fusion for Precise Brain Tumour Segmentation
Ananya Gupta (1) MARS ROVER MANIPAL
  • UATTA-ENS: Uncertainty Aware Test Time Augmented Ensemble for PIRC Diabetic Retinopathy Detection
Saurabh Kumar Mishra (1) MARS ROVER MANIPAL
  • UATTA-ENS: Uncertainty Aware Test Time Augmented Ensemble for PIRC Diabetic Retinopathy Detection
Mihir Agarwal (1) RESEARCH SOCIETY MANIPAL
  • UATTA-EB: Uncertainty-Aware Test-Time Augmented Ensemble of BERTs for Classifying Common Mental Illnesses on Social Media Posts
Danush Khanna (1) MANIPAL UNIVERSITY JAIPUR
  • SELF-PERCEPT: Introspection Improves Large Language Models' Detection of Multi-Person Mental Manipulation in Conversations
Sidhaarth Sredharan Murali (1) NIT SURATHKAL
  • SELF-PERCEPT: Introspection Improves Large Language Models' Detection of Multi-Person Mental Manipulation in Conversations
Aditya Kumar Guru (1) MANIPAL UNIVERSITY JAIPUR
  • SELF-PERCEPT: Introspection Improves Large Language Models' Detection of Multi-Person Mental Manipulation in Conversations
Siddharth Shukla (1) MANIPAL UNIVERSITY JAIPUR
  • SELF-PERCEPT: Introspection Improves Large Language Models' Detection of Multi-Person Mental Manipulation in Conversations
Tanuj Tyagi (1) MANIPAL UNIVERSITY JAIPUR
  • SELF-PERCEPT: Introspection Improves Large Language Models' Detection of Multi-Person Mental Manipulation in Conversations
Sandeep Chaurasia (1) MANIPAL UNIVERSITY JAIPUR
  • SELF-PERCEPT: Introspection Improves Large Language Models' Detection of Multi-Person Mental Manipulation in Conversations
Kripabandhu Ghosh (1) IISER KOLKATA
  • SELF-PERCEPT: Introspection Improves Large Language Models' Detection of Multi-Person Mental Manipulation in Conversations
Krish Didwania (1) MARS ROVER MANIPAL
  • AgriLLM: Harnessing transformers for farmer queries.
Amit Agarwal (1) WELLS FARGO AI COE
  • AgriLLM: Harnessing transformers for farmer queries.
Rashi Goel (1) RESEARCH SOCIETY MANIPAL
  • RSM-NLP at BLP-2023 Task 2: Bangla Sentiment Analysis using Weighted and Majority Voted Fine-Tuned Transformers
Komal Mathur (1) RESEARCH SOCIETY MANIPAL
  • RSM-NLP at BLP-2023 Task 2: Bangla Sentiment Analysis using Weighted and Majority Voted Fine-Tuned Transformers
Swetha Vemulapalli (1) RESEARCH SOCIETY MANIPAL
  • RSM-NLP at BLP-2023 Task 2: Bangla Sentiment Analysis using Weighted and Majority Voted Fine-Tuned Transformers
Hemang Malik (1) RESEARCH SOCIETY MANIPAL
  • HGP-NLP at Shared Task: Leveraging LoRA for Lay Summarization of Biomedical Research Articles using Seq2Seq Transformers
Gaurav Pradeep (1) RESEARCH SOCIETY MANIPAL
  • HGP-NLP at Shared Task: Leveraging LoRA for Lay Summarization of Biomedical Research Articles using Seq2Seq Transformers