Data Scientist Machine Learning

Rajveer Shringi

SENiOR DATA SCiENTiST · MEDiAMARKTSATURN TECH GMBH

**, *********** ****** *****, *******

+49-176-******** ***.*******@*****.*** dupree rajveer-shringi Rajveer Shringi Summary

An Experienced Data Scientist at MediaMarktSaturn Technology GmbH, a leading retail electronics supplier, specializing in developing cutting edge Search and Recommendation technologies for e commerce. Leveraging NLP and LLMs to enhance query understanding and semantic search, minimizing zero result searches and contributing to a more engaging customer experience. Leading to significant improvements to key product performance indicators, including a 60 80% reduction in session exit rates. ExpertiseencompassesSearchandRecommendationsystems,Personalization, Ranking, aswellasNaturalLanguageProcessing/Understanding, conversational intelligence, and voice enabled platforms. Demonstrated expertise in building, deploying, and operationalizing machine learning systems end to end, leveraging MLOps principles and deploying robust, scalable endpoints on AWS and GCP. Integration in A/B testing and creating dashboards for continuous model monitoring and drift analysis.

Recipient of the DAAD Stibet 2017 scholarship for outstanding academic potential at TU Kaiserslautern Technical skills & competences

Programming Languages python, typescript, c/c++ with boost, java Machine Learning specific tensorflow( gpu ), pytorch, pyspark, conda, jupyter notebook, pandas, scikit learn. NLU & Search specific finetuning and customizing large language models LLMs, vector seach, sequence tagging, intent classification deep neural embeddings, spacy, nltk, gensim.

Cloud Google Cloud Platform, Vertex AI, Amazon Web Service (AWS), Sagemaker Database and Servers mySQL, postgreSQL

Enterprise Software Development Jira : for Agile Project Management, github: version control Education

Technische Universität Kaiserslautern Germany

M.SC. iN INFORMATiCS (SPECiALiZATiON: INTELLiGENT SYSTEMS AND SOFTWARE ENGiNEERiNG) Oct. 2015 Nov. 2018

• CGPA: 1.28/5 (1: VeryGood, 5: Insufficient)

Courses included Machine Learning (AI Applications, Deep Neural Networks), Recommender Systems, and Multimedia Social Web Mining.

• Master’s Thesis: Recommender System using Semantic Document Similarity

– Developed a recommender system for books in German for a leading book seller in Germany in collaboration with Thalia & SAP SE.

– Employed statistical techniques (e.g., TF/IDF, Topic Modeling) and semantic methods (e.g., word and document embeddings).

• Project: AirScript Gesture based Document Creation

– Executed in partnership with the German Research Center for Artificial Intelligence (DFKI, Kaiserslautern).

– The results were published at the 14th IAPR International Conference on Document Analysis and Recognition (ICDAR 2017). MIT, Manipal University India

B.E. iN COMPUTER SCiENCE AND ENGiNEERiNG July 2009 Mar. 2013

• CGPA: 7.62/10 (10: VeryGood, 5: Insufficient)

• Bachelor’s Thesis: Automation Testing Framework

• Developed an automation testing framework with the Product Automation team at RSA EMC Corporation. Publication(s)

ICDAR 2017 In collaboration with DFKI, Kaiserslautern. AirScript: Creating documents in air. https://arxiv.org/abs/1705.11181 TU KL 2017 Introduction to Blockchain Technology. https://github.com/dupree/blockchain_paper JAN, 2024 SHRiNGi, RAJVEER · RESUME 1

Employment History

MediaMarktSaturn Technology GmbH Berlin, Germany

SENiOR DATA SCiENTiST December 2022 Current

• Developing innovativeSearchandRecommendationtechnologyfore commerce,leveragingNLPandLLMstoenhancequeryunderstanding and reduce zero result occurrences, improving customer engagement and experience.

• Implementing semantic search solutions, optimizing search results by training and fine tuning LLMs on domain specific data for superior relevance and accuracy. These efforts led to significant improvements in key product performance indicators, notably a 60–80% reduction in session exit rates.

• Collaborating cross functionally with product owners, search managers, campaign, and brand managers to define use cases, propose fea tures, and drive implementation from ideation to delivery.

• Building, deploying, andoperationalizingmachinelearningsystemsend to endusingMLOpsprinciples,deployingscalableendpointsoncloud GCP, and developing dashboards for continuous monitoring and drift analysis. StepStone GmbH Berlin, Germany

SENiOR MACHiNE LEARNiNG ENGiNEER Nov 2021 Mar 2022

• Developed conversational systems and agent based solutions for Job Seeker Assistance technology Mya.

• Implemented dialogue management, NLU (intent classification, entity recognition), and recommendation serving via chatbot interfaces to collect user feedback.

• Collaborated with product team to define use cases and translate them into actionable technical specifications for agent based systems. Berlin, Germany

NLP ENGiNEER Aug 2020 Nov 2021

• Implemented a learning to rank algorithm to match job candidates with relevant job listings, leveraging supervised ML methods such as LightGBM with domain specific features.

• Utilized Shapley values for model explainability, enhancing transparency in candidate job matching.

• Contributed to search, match, and recommendation systems in the recruitment space, applying state of the art methods including deep neural embeddings for improved performance.

German Autolabs GmbH Berlin, Germany

NLP ENGiNEER Aug 2019 Aug 2020

• Developed core components of a conversational (chatbot) system for an automotive voice assistant, including dialogue management, NLU

(intent classification, entity recognition), and speech recognition.

• Collaborated cross functionally with product development, design, and customer success teams to ensure seamless integration and enhance user experience.

SAP SE Wiesloch Walldorf, Germany

INTERNSHiP MASTER THESiS STUDENT RESEARCH ASSiSTANT March 2018 Jan. 2019

• Developed NLP solutions using classical and deep neural network methods for text classification, clustering, and feature extraction.

• Developed a content based Recommender System using semantic document similarity for a leading German bookseller.

• Worked with SAP Healthcare dataset for medical record classification and Named Entity Recognition (NER) using TensorFlow. Fraunhofer Research Institute, ITWM Kaiserslautern Berlin, Germany SCiENTiFiC RESEARCH EMPLOYEE(FULL) April Aug 2019

• Research and Development Engineer with the High Performance Computing group.

• Contributed to the GreenByIT renewable energy project, optimizing charging and discharging cycles for renewables powered battery using time series forecasting with LSTM/RNN and related forecasting models. College Park Maryland, USA

SEMESTER ABROAD STUDENT RESEARCH ASSiSTANT Sep. 2017 Dec. 2017 Kaiserslautern, Germany

SOFTWAREDEVELOPER(PART TiME) Feb. 2016 Jan. 2019

• Designed and developed a full stack web application for interfacing with the Finite Pointset Method (FPM), a grid free numerical simulation tool. Frontend C++ Wt widgets, Backend MySql

JAN, 2024 SHRiNGi, RAJVEER · RESUME 2

Altair India Engineering Ltd, EMC Corporation Bangalore, India SOFTWARE ENGiNEER Jan 2013 June 2015

• Full stack development using Java and Python, contributing to flagship products such as HyperWorks, HyperMesh, and EMC Healthcare Integration Portfolio.

Projects

Search & Learning to Rank Talent and Job Match Berlin, Germany SELF LEARNiNG AND PROTOTYPiNG April 2024

• Matching talents with relevant job opportunities based on attributes like degree, job roles, salary, and languages.

• Data cleaning, transformation, and feature extraction were performed to prepare the dataset

• A logistic regression model using scikit learn to classify talent job matches and generate a ranking score.

• The model outputs a label (true/false for match) and a score for ranking the results.

• Emphasis placed on clean, well structured Python code and best practices.

• GitHub repository url

Data Augmentation Berlin, Germany

SELF LEARNiNG AND PROTOTYPiNG Feb 2019

• Given the list of job offers (title, description and seniority level), some of the offers are missing the seniority level.

• Augmenting the missing information using inferences from available data

• Text classification, semantic similarity, tfIdf, word movers distance

• GitHub repository url

AirScript Creating documents in Air KL, Germany

DFKI, KAiSERSLAUTERN Oct. 2016 Feb. 2017

• Gesture recognition using time series data analysis from IMU’s mounted on an armband.

• Resulting paper published at 14th IAPR International Conference on Document Analysis and Recognition (ICDAR 2017).

• https://arxiv.org/abs/1705.11181

Language skills

English Near native proficiency C2 TOEFL score 115/120 German Intermediate B2.1

Hindi Mother tongue

Awards and Appreciations

• Recipient of DAAD Stibet 2017 student scholarship for promising academic candidates at TU Kaiserslautern JAN, 2024 SHRiNGi, RAJVEER · RESUME 3

Contact this candidate