Prakhar Verma

I am a Senior Machine Learning Engineer at Inven, where I build reliable, production-scale LLM systems. My work focuses on grounding language models in private, multimodal data, building scalable data pipelines for AI systems, and applying probabilistic machine learning for well-calibrated forecasting and estimation.

I completed my Ph.D. at Aalto University, Finland, under the supervision of Prof. Arno Solin. My experience spans both industry and academia through research roles at Microsoft Research, Adobe Research, and the University of Oxford, with publications at ICML (Oral), NeurIPS, AISTATS, and TMLR.

I am particularly interested in the intersection of probabilistic modeling and agentic systems, with the goal of developing principled, uncertainty-aware agent frameworks grounded in Bayesian principles for evaluation, failure detection and recovery, and decision-making under uncertainty.

News

Aug 2026 Defended my Ph.D. dissertation Scalable Probabilistic Inference for Sequential Stochastic Models at Aalto University, Finland advised by Prof. Arno Solin and examined by Prof. Jes Frellsen (DTU).
Apr 2026 Prakhar Verma, David Arbour, Sunav Choudhary, Harshita Chopra, Arno Solin, Atanu R. Sinha. Sequential Causal Discovery with Noisy Language Model Priors accepted in Transactions on Machine Learning Research (TMLR).
Oct 2025 Started as a Senior Machine Learning Engineer at Inven.
Aug 2025 I will be a workflow chair for AISTATS 2026 with Yo Joong Choe and Program Chairs Aaditya Ramdas and Arno Solin.
Mar 2025 Prakhar Verma, Sukruta Prakash Midigeshi, Gaurav Sinha, Arno Solin, Nagarajan Natarajan, Amit Sharma. Plan*RAG: Efficient Test-Time Planning for Retrieval Augmented Generation accepted in Workshop on Reasoning and Planning for Large Language Models at ICLR 2025.

Selected Publications

See full list on Google Scholar.

Sequential Causal Discovery paper thumbnail

Sequential Causal Discovery with Noisy Language Model Priors

Prakhar Verma, David Arbour, Sunav Choudhary, Harshita Chopra, Arno Solin, Atanu R. Sinha

Transactions on Machine Learning Research (TMLR), 2026

OpenReview arXiv

Plan*RAG: Efficient Test-Time Planning for Retrieval Augmented Generation

Prakhar Verma, Sukruta Prakash Midigeshi, Gaurav Sinha, Arno Solin, Nagarajan Natarajan, Amit Sharma

Workshop on Reasoning and Planning for Large Language Models, ICLR 2025

OpenReview arXiv Poster

Variational Gaussian Process Diffusion Processes

Prakhar Verma, Vincent Adam, Arno Solin

International Conference on Artificial Intelligence and Statistics (AISTATS), 2024

PMLR arXiv Code Poster

Memory-based Dual Gaussian Processes paper thumbnail

Memory-based Dual Gaussian Processes for Sequential Learning

Paul E. Chang^*, Prakhar Verma^*, ST John, Arno Solin, Mohammad Emtiyaz Khan

International Conference on Machine Learning (ICML), 2023 Oral

PMLR arXiv Code Poster

Sequential Learning in GPs paper thumbnail

Sequential Learning in GPs with Memory and Bayesian Leverage Score

Prakhar Verma, Paul E. Chang, Arno Solin, Mohammad Emtiyaz Khan

Continual Lifelong Learning Workshop, ACML 2022 Contributed Talk

OpenReview Code Slides

Scalable Inference in SDEs paper thumbnail

Scalable Inference in SDEs by Direct Matching of the Fokker–Planck–Kolmogorov Equation

Arno Solin, Ella Tamir, Prakhar Verma

Advances in Neural Information Processing Systems (NeurIPS), 2021

OpenReview arXiv Code Poster

Theses

Prakhar Verma (2026). Scalable Probabilistic Inference for Sequential Stochastic Models. Doctoral dissertation, Department of Computer Science, Aalto University, Finland.

Prakhar Verma (2021). Sparse Gaussian Processes for Stochastic Differential Equations. Master's thesis, Department of Computer Science, Aalto University, Finland.

Prakhar Verma (2016). Development of Automated GIS Tools on Various Platforms. Bachelor's thesis. Uttarakhand Technical University, India. In collaboration with TomTom India.