Sandeep Samant Email & Phone Number

Bengaluru, Karnataka, India

Bengaluru, Karnataka, India

Bengaluru, Karnataka, India

Aachen, North Rhine-Westphalia, Germany

• 1) Working on Prompt based Chat bot using Langchain framework which leverages Open AILLM (GPT 3) API in the backend2) Contributed to build a custom RL agent using DDPG algorithm to train model for 100,000 episodes and.
• Achieved an accuracy rate of 95% in validating the model results against ground truth data.4) Developed an optimization model using operations research (OR) techniques to address thechallenge of carbon emissions from.
• Achieved an average cost reduction of 25% per installation, resulting in significant cost savingsfor homeowners and a reduction in CO2 emissions of over 500,000 tons annually.Tech Stack: Python, CausalNex, CausalML.

Essen, North Rhine-Westphalia, Germany

• 1) Led a project to centralize customer and their transaction data from Virta API for 5 years, covering over 1 million customers.
• Developed an automated data pipeline with parallelization using Azure Data Hub, improving data handling efficiency by 30%.
• Consolidated customer and transaction data, gaining valuable insights into behavior over timeand driving a 20% increase in customer acquisition and a 15% increase in customer retention. 2) Designed and developed Power.

Aachen, North Rhine-Westphalia, Germany

Implemented ReactJs functionalities like improving the existing web interface for hardware products.

Pune, Maharashtra

• 1) Built a migration component to move all compliance data (10,000 users) from SQL DB server to foundry.2) Developed an AI agent to help answer questions about general compliance regulations
• Trained an NLP Q&A retriever model (based on BM25 algortihm in Haystack) on a dataset of compliance cases, achieving an accuracy of over 90% on a held-out validation set
• Improved the accuracy of the existing baseline model by 15% resulting in a significant reduction in the time required to answer routine compliance queries, with an estimated 30% reduction in response time and a 20%.

Pune, Maharashtra, India

• Built a capacity management ML model to automate KYC case assignment to an individual based on Historical data resulting in 50% reduction in manual task
• Contributed to development and testing of the backend code for a mission-critical application called "Customer Liquidity Management".
• Addressed the latency issues in API calls by optimizing the payload structure, resulting in a 20% improvement in overall speed and performance.Tech Stack: Python, Sklearn, SHAP, Flask, Javascript, HTML, CSS, SQL, AWS

Pune, Maharashtra, India

1) Worked on 2 PoCs which later became a part of one of the products. 2) Built a bidirectional data schema transform layer between the legacy RDBMS and the new document store based on MongoDB in NodeJS.

Work From Home (@Indian Mistry)

Internship to develop the Profile app of the company.

M.Sc. Data Science, 1.8 ( 1- Maximum, 5 - Fail ) - Made It To The Dean List

Master Thesis: Novelty Detection in NILM using Machine Learning. (Grade: 1.0) Took Courses : Mathematics of Data Science Machine Learning.

Bachelor Of Engineering (B.E.), Computer Science, 9.2 Sgpa

Activities and Societies: IIH CLUB, OSS CLUB,ALUMNI MEET CLUB, PACE ORGANISING COMMITEE, ALCHEMY ORGANISING COMMITEE.An active member of.

Xii, Mathematics And Computer Science, A+

Was among meritorious students of the school, took part in many national level competitions.

High School, Mathematics, A+

Was the head boy of the school, took part in many public speaking events and other activities.