Available Now • Stage de fin d'études • Paris, France

Engineering
Intelligence.

Production data engineer with four years at Accenture building large-scale ML analytics systems for banking, aviation, and consumer electronics clients. Now completing an MSc at KEDGE Business School, specialising in applied machine learning and generative AI.

Deepayan Sarkar

Enterprise Engineering

Accenture

2021—2025 • Kolkata, India

Software Engineer

Sep 2023 — Aug 2025

Engineered robust Python/SQL data pipelines ensuring zero data loss across production workloads. Delivered 12+ end-to-end analytics projects on GCP on time and within budget by coordinating across 15+ cross-functional stakeholders.

0M+
Daily Records Processed
0%
ETL Time Reduction
0+
Hrs Saved Monthly

Associate Software Engineer

Aug 2021 — Sep 2023

Maintained critical GCP analytics infrastructure (BigQuery, Cloud Storage, Dataflow) supporting business-critical workloads. Analysed large-scale datasets (1M+ rows) to surface trends and anomalies guiding product and marketing decisions.

0%
GCP Infrastructure Uptime
0%
Lift in Customer Experience via A/B Testing

Professional Validation

“Deepayan is a highly capable and analytical problem-solver. What stands out to me is his ability to take high-level guidance and translate it into rigorous research, confidently navigating challenges like label scarcity through solid statistical validation.”
Tianyuan Zhang

Tianyuan Zhang

Assistant Professor, KEDGE Business School

“Deepayan took ownership of key deliverables easily. He is a smart and driven individual who consistently explores innovative approaches to problem solving and brings his own unique perspective to the team. He is a quick learner, always eager to pick up new technologies.”
Taqabul Nisha

Taqabul Nisha

Digital Analytics Lead at e& UAE

Applied Projects & Hackathons

BNP Paribas Hackathon · KEDGE

AI Persona Bots

Multi-agent platform synthesising 2,438 synthetic customer responses across 8 segments for credit product marketing research.

0%
Coherence
88.6%
Relevance
<1hr
vs 4–8wks
Azure AI FoundryLangChainGPT-4o
L'Oréal Hackathon · KEDGE

Skincare Product Classifier

Multi-label text classifier using LinearSVC and One-vs-Rest across 33 beauty categories, with per-class threshold optimisation for minority-class recall.

0.67
F1 Score
6,240
Products
~18%
Metric Lift
Pythonscikit-learnTF-IDFNLP
View on GitHub ↗
Personal Project · Android

Offline Vision Assistant

Fully offline Android app for the visually impaired — on-device object detection, voice feedback, and CI/CD pipeline via GitHub Actions.

<200ms
Inference
100%
Offline
TensorFlow LiteFlutterDartOn-Device ML
View on GitHub ↗
Tableau Public · Data Viz

China Trade Transport Analysis

Interactive dashboard exploring China's import/export transport patterns — shipping modes, commodity flows, and YoY growth across global corridors.

4
Transport Modes
YoY
Growth Analysis
TableauLOD ExpressionsData Viz
View on Tableau Public ↗

Technical Specifications

GenAI & Machine Learning

  • LLMs, RAG & Multi-Agent Systems
  • LangChain & Azure AI Foundry
  • HuggingFace Transformers
  • scikit-learn & XGBoost
  • PyTorch & TensorFlow

Cloud & Data Engineering

  • GCP (BigQuery, Dataflow, Cloud Storage)
  • Airflow & ETL Pipelines
  • Docker & CI/CD
  • MLOps Architecture

Analytics & Programming

  • Python, SQL & Apache Spark
  • Power BI & Tableau
  • A/B Testing & Statistical Modelling
  • Exploratory Data Analysis

Academic & Credentials

Sep 2025 — APR 2027 • Bordeaux, France

KEDGE Business School

MSc in Data Analytics for Business

Eligible for Convention de Stage
French Work Authorisation
Jul 2017 — May 2021 • Kolkata, India

University of Engineering & Management

BEng in Electronics & Communication Engineering

Certifications

  • Google Cloud CertifiedAssociate Cloud Engineer (2023)
  • Microsoft Azure AI FundamentalsAI-900 (2026)
  • Professional Scrum Master 1Scrum.org (2026)
  • SQL for Data ScienceCoursera / UC Davis (2022)