Open to opportunities

Vikrant Sahu

@vikrantsahu

Message

Senior Data Scientist specializing in ML, NLP, and scalable end-to-end data solutions.

India

Message

What I'm looking for

I seek roles focused on building production-grade ML/NLP systems, mentorship, and measurable business impact in collaborative, data-driven teams.

I am a Senior Data Scientist with eight years of experience across investment banking, fintech, pharmaceuticals, and startups, turning data into actionable insights and measurable business impact.

I have led teams and projects—building production ML pipelines with PySpark/Databricks, designing multiheaded LSTM architectures, and delivering a Customer Lifetime Value model with 7–10% MAPE that generated ~5% incremental revenue and ~10% lower CAC.

My hands-on work includes developing RAG chatbots using open-source LLMs, establishing vector DB retrieval systems, engineering BERT classification frameworks with PyTorch/Hugging Face, and deploying models with CI/CD using Flask and Docker.

I prioritize model transparency and monitoring (SHAP, MLflow), pragmatic deployment, and mentoring—having guided interns and run trainings—while driving measurable outcomes like reduced ticket resolution time, improved retention, and faster model development.

Experience

Work history, roles, and key accomplishments

Current

Data Scientist - 3

Current

AngelOne

Apr 2024 - Present (1 year 6 months)

Led a team of 3 on a Customer Lifetime Value project, achieving MAPE of 7–10%, generating ~5% incremental revenue and reducing CAC by ~10%. Designed a multi-headed LSTM and built an end-to-end PySpark pipeline on Databricks for feature engineering, training, and updates.

Pyspark Databricks Time Series

Senior Data Scientist

Concentrix Global Services

Mar 2023 - Mar 2024 (1 year)

Developed a RAG chatbot using open-source LLMs and LangChain achieving 90% efficiency across 5,000+ documents, and built a vector DB retrieval system with high-accuracy ensemble retrieval methods.

RAG Lang Chain Retrieval Methods FastChat NLP

Data Scientist

HSBC Global Research

Mar 2021 - Mar 2023 (2 years)

Engineered an end-to-end BERT classification framework reducing modelling time by 70% and created an ML asset-allocation strategy with 73–76% accuracy; implemented monitoring and retraining workflows using MLflow.

BERT PyTorch Hugging Face SHAP MLFlow NLP Model Monitoring

Data Scientist

Near.Co

Nov 2020 - Mar 2021 (4 months)

Built PySpark big-data pipelines to analyze terabytes of ads/location data and derived footfall and visitor-attribute insights for retail units globally.

Pyspark Big Data Ad Insights Data Pipelines Amazon SageMaker

Data Scientist

Digite Inc.

May 2018 - Oct 2020 (2 years 5 months)

Applied NLP to improve ticket resolution time by 40% and efficiency by 60%, and developed a multi-field recommendation system reducing form fill time by 50%; productionized ML models with CI/CD.

NLP Python Flask Docker CI CD Hugging Face