Camille's Blog

IMT ATLANTIQUE, Engineering Diploma

2020-2022

Brest, France

Relevant Courseworks:

• Optimized Deep Learning (MicroNet challenge : regularization, pruning and quantization techniques to reduce the floating point number and number of parameters)
• Machine Learning (unsupervised (clustering, PCA/EFA) & supervised (Regression: prediction of the number of insults on a Youtube Video – Classification: image classification with ResNet))
• Advanced Big Data (Cassandra, Hadoop, Spark, Kafka, Storm)
• Object Oriented Programming (Transcripen project for speech-to-text analysis)
• Advanced Mathematics
• Network (OSI model, DNS, QoS)

OTTO-FRIEDRICH-UNIVERSITÄT BAMBERG, Erasmus Program

April-July 2019

Bamberg, Germany

Relevant Courseworks:

• International Finance
• European and International Economics
• Economics of Inequality

UNIVERSITY OF CAMBRIDGE, Exchange Program

September-December 2018

Cambridge, UK

Relevant Courseworks:

• Corporate Finance
• Entrepreneurial Finance
• Economics

GRENOBLE ECOLE DE MANAGEMENT, Grande Ecole Diploma

2017-2021

Grenoble, France

Relevant Courseworks:

• Corporate Finance
• Thesis on “Improving Stock Return prediction’s traditional Algorithm of Fama & French thanks to recurrent neural networks”
• Economics
• Law
• HRM
• Management

Pierre De Fermat, Classe Préparatoire aux Grandes Ecoles (ECS)

2015-2017

Toulouse, France

Relevant Courseworks:

• Maths
• Geopolitics

ORANGE, Machine Learning Engineer

2022-now

Paris, France

NLP Machine Learning Engineer within a Research Project about LLM adaptation to the telecommunication industry which aims at getting a "small" (7b-parameters) model performing as well as big LLMs on telecom use-cases:

• Explored best practices about fine-tuning for domain adaptation of LLMs : on the data, preprocessing, continual pretraining (self-supervised fine-tuning on raw domain texts), instruction-tuning (on telco instructions (QA, MCQs, Summarization)), fine-tuning process itself (packing vs non-packing, loss computed on all tokens (auto-regressive way) vs on output tokens only, ...), parameter-efficient fine-tuning methods (lora vs qlora vs prefix-tuning vs full fine-tuning)
• Implemented a pipeline for domain-adaptation of LLM (from data collection, preprocessing, to fine-tuning, to results evaluation)
• Created an internal blog about LLM to help collaborators fine-tuning / using LLMs at inference
• Trained models on a SLURM internal cluster using several parallelization methods
• Wrote a research paper on our findings about domain-adaptation

Development of an internal tool for data ingestion and processing

Development and deployment of a crawler using scrapy to retrieve PDFs from difference telco websites (3GPP, 5GAA, ETSI, ...) and internal sources

Implementation and comparison of different PDF processing methods to retrieve texts and tables (Tesseract, Nougat, PDFMiner, ...)

DELPHA.IO, Machine Learning Engineer

2021-2022

Paris, France

Developed and deployed AI models for data cleaning of Salesforce's customer databases

• Developed an algorithm to perform large-scale string comparison via an optimisation of sparse matrix multiplication (vectorization of names + cosine similarity computation) to compare clients names from Salesforce environments databases in order to detect duplicates client names into their database as well as linking client names from the database which have a specific relationship (parent/subsidiary relationship, same group, …)
• Deployment of algorithms developed on AWS: use of Appflows to get/send back the data (storage in S3), AWS EMR to make an initialisation process running, AWS lambda for batch processes, API endpoints construction for Salesforce, creation of AWS stacks on CloudFormation
• Product owner: worked on an Agile way (SCRUM) to define each task / the timeline / the difficulties on both projects described above

ING BANK, Leveraged Finance intern

January-June 2020

Paris, France

Prospected Private Equity funds: Analyzing the LBO-market news by realizing daily Press Reviews for the team, gathered information of potential LBO targets within a mapping indicating which funds is interested in which potential LBO target, realized ‘one-pagers’ for potential interesting companies which are sold/expected to be sold to give funds LBO ideas

Pitched funds on potential LBOs with business and market descriptions, financial analysis (P&L, Cashflow Statement, Budget, deleveraging profile, current trading and exit considerations analyses), potential upsides proposals, key investors’ concerns and key credit considerations

Built LBO models by challenging the Business Plan given by the management/sponsor by analyzing the assumptions taken

Took part in the Credit Modification process for ‘Prêts Garantis par l’Etat’ (PGE) and waivers (change of covenants’ rules in critical situations) assignations during the Covid-19 crisis

BNP PARIBAS, Leveraged Finance intern

July-December 2019

Paris, France

Pitched funds on potential LBOs (new LBO transaction, refinancing, extension, restructuring)

Analyzed LBO targets (comps, industry research, identified risks and mitigants, analyzed financials) and built LBO models

Took part in the Syndication process: Lender Presentations and Rating Presentations

AQUA ASSET MANAGEMENT, Front Office Intern in Private Equity and Asset Management

December 2018-March 2019

Paris, France

Conducted industry research to find investment opportunities in German countries aligning the fund’s investment strategy through Factset and prepared business/financial presentations of potential targets (teasers) for my managers

FIDSUD AUDIT, Financial auditor intern

May-July 2018

Toulouse, France

Assisted in auditing 16 firms by identifying missing information, ambiguities in reports and resolved them accordingly

Languages

Native C2 B2

Programming

Advanced Intermediate Basis

AI Libraries

Pytorch: Advanced, Tensorflow: Intermediate, Keras: Basis

Transformers: Advanced, Langchain: Intermediate

Dev Skills

Advanced Intermediate Advanced

Big Data tools / languages

Basis, Intermediate

Cloud Skills

AWS: Intermediate, GCP: Basis

HACKATLANTIQUE, CTF attendent

March 2022

Rennes, France

Attended the hacking championship organized by my school (specific piece of text to find hidden on the server / behind webpages, …)

COMU, Dancer

2017-2018

Grenoble, France

Dancer at a Musical from Grenoble Ecole de Management (Cosa-Nostra)