About Experience Publications Contact
Technical Lead · AI Operations · NLP

Sanjay
Kamath

PhD in NLP from Université Paris-Saclay. Technical Lead — AI Operations (AIOps) at TotalEnergies Digital IT. Based in Paris, France.

Sanjay Kamath

Background & Skills

I defended my PhD thesis on
"Question Answering with Hybrid Data and Models"
on 6 February 2020 at LIMSI (Université Paris-Saclay).

Currently Technical Lead on the AI Operations (AIOps) team
at TotalEnergies Digital IT,
leading NLP and AI initiatives across the organisation.

Read my thesis →    PhD presentation slides →

Research Interests

Natural Language Processing Question Answering Deep Learning Machine Learning Text Analysis Biomedical NLP

Technical Stack

LLMs & AI
LLM Fine-tuning RAG Transformers BERT / GPT / LLaMA Hugging Face
LLM Serving & MLOps
MLflow model quantization
Cloud & Infrastructure
Azure AWS Docker
Languages & Frameworks
Python PyTorch Scikit-Learn

Education

PhD — Computer Science (NLP)
Université Paris-Saclay  ·  2016–2019
MSc — AI & Web
ENSIMAG / Université Grenoble Alpes  ·  2014–2016
BE — Computer Science
Visvesvaraya Technological University  ·  2010–2014

Work Experience

Technical Lead — AI Operations (AIOps)
Apr 2026 – Present
TotalEnergies · TGITS Digital IT · Paris, On-site
Technical Lead for the Artificial Intelligence Operations squad within the Digital IT department of TGITS.
Senior Data Scientist
Jul 2023 – Apr 2026
TotalEnergies · TGITS Métiers Technique (MTE) · Greater Paris, Hybrid
Senior Data Scientist in the Data Support Squad, managing Data Science projects and mentoring junior scientists for domain experts at the Exploration branch across different geographical hubs.

Projects included: data extraction from legacy reports using computer vision, ML modelling for petrophysical property prediction, generative NLP models for data extraction, and migration to low-code/no-code platforms.
Research Scientist — R&D Power Group
Nov 2019 – Jun 2023
TotalEnergies · Greater Paris
Research on Machine Learning on textual data, Applied NLP, and Information Retrieval.
Multi-modal models combining linguistic knowledge for image captioning and detection.
Domain-specific model training on High Performance Computing (HPC) architectures.
Intelligent document processing using Image Processing and AI tools.
Built prototypes and demos to demonstrate POCs; coordinated with external partners.
Doctoral Research
Oct 2016 – Oct 2019
LIMSI / LRI · Université Paris-Saclay
Topic: Question Answering with hybrid data and models.
ANR Project GOASQ.
Supervisors: Prof. Brigitte Grau & Assoc. Prof. Yue Ma.
Teaching Assistant
Jan 2017 – May 2019
IUT d'Orsay
Project supervision: Android, Unity 3D & Garmin Wear (semester 4).
Practical sessions: Human–Machine Interaction (semester 2).
Research Intern (Master 2)
Nov 2015 – Jun 2016
LIG Grenoble
Master thesis under Prof. Marie-Christine Fauvet & Prof. Lorraine Geouriot.
Research Intern (Master 1)
Feb 2015 – Aug 2015
LIG Grenoble
Research internship under supervision of Prof. Marie-Christine Fauvet.

Publications

Measuring semantic similarity of clinical trial outcomes using deep pre-trained language representations
JBMI · Oct 2019
Journal of Biomedical Informatics: X
Paper ↗
How to Pre-Train Your Model? Comparison of Different Pre-Training Models for Biomedical Question Answering
BioASQ · ECMLPKDD 2019
7th BioASQ Workshop, September 2019
Paper ↗
Predicting and Integrating Expected Answer Types into a Simple Recurrent Neural Network Model for Answer Sentence Selection
CICLING 2019
20th International Conference on Computational Linguistics and Intelligent Text Processing
Paper ↗
An Adaption of BIOASQ QA Dataset for Machine Reading Systems by Manual Annotations of Answer Spans
BioASQ · EMNLP 2018
6th BioASQ Workshop, October 2018
Paper ↗
Verification of the Expected Answer Type for Biomedical Question Answering
WWW 2018
HQA Workshop at The Web Conference, April 2018
Paper ↗
A Study of Word Embeddings for Biomedical Question Answering
SIIM 2017
4e édition du Symposium sur l'Ingénierie de l'Information Médicale
Paper ↗
Offering Context-Aware Personalised Services for Mobile Users
ICSOC 2015
International Conference on Service-Oriented Computing
Paper ↗
Processing NL queries to disambiguate named entities and extract users' goals: application to e-Tourism
RJCRI CORIA 2015
Junior Researchers track at the French IR Conference, Toulouse
Paper ↗
Using Explicit and Implicit Feedback to Model Context for Proactive Recommendations
LIG Grenoble 2016
Master 2 Research Internship Report
Report ↗
Discovering Context Information and Parameters from a Natural Language Query
LIG Grenoble 2015
Master 1 Research Internship Report
Report ↗

Languages

English
Fluent
French
Professional
Kannada
Native
Konkani
Native
Hindi
Professional

Contact