Yogesh Shankariya
Ahmedabad
Summary
Data Science professional with over 3 years of experience, specializing in the development and deployment of advanced analytics solutions. Honored as a part of the prestigious Infosys Platinum Club, ranking within the top 1% of data and analytics for technical prowess and performance. Awarded the Insta and Rise awards for exceptional performance.
Experienced team leader, managing a team of two, and successfully led the creation of six proof-of-concepts (PoCs), two of which have been implemented into production for fraud detection and contact center intelligence. These PoCs spanned various domains including complaint analysis, automated letter generation, chatbot development, contact center intelligence, fraud detection, and financial crisis detection, showcasing proficiency in a wide range of data science tools and platforms.
Overview
3
3
years of professional experience
1
1
Certification
Work History
Lead Data Scientist
Infosys Ltd.
03.2024 - Current
- Built a proof-of-concept (PoC) for complaint analysis on Azure.
- Used GPT-3.5 to summarize complaints registered by customers, identify causes, and assess impacts.
- Applied OpenAI’s embedding for clustering similar issues, and assigned generic labels for easy interpretation.
- Visualized insights via PowerBI, aiding in understanding customer issues and agent performance.
- The PoC effectively showcased AI’s potential in improving customer service processes.
Lead Data Scientist
Infosys Ltd.
03.2024 - Current
- Built a proof-of-concept (PoC) for automated rectification letter generation using OpenAI’s GPT-4 Turbo.
- Used few-shot prompt engineering to generate letters based on outcomes: upheld, refer, or reject.
- Employed different templates for few-shot learning.
- Produced final letters based on the outcome, rectification amount, and customer issues.
- The PoC effectively automated letter writing, saving agents’ time and effort.
Lead Data Scientist
Infosys Ltd.
03.2024 - Current
- Developed a proof-of-concept (PoC) for a RAG (Retrieve and Generate) based chatbot.
- Provided users with the flexibility to upload PDFs and DOCs, enhancing the chatbot’s versatility.
- Offered users the option to select from various models including OpenAI’s GPT-4.0 and the local Llama3 model, catering to diverse user needs.
- Built the chatbot based on LangChain and utilized Streamlit for UI development, ensuring a user-friendly interface.
Lead Data Scientist
Infosys Ltd.
09.2023 - 03.2024
- Built an AWS MLOps pipeline for Contact Center Intelligence, converting voice to text using FFmpeg and AWS Transcribe.
- Ensured data privacy by implementing PII removal and data masking, and prepared data for machine learning with a Bag of Words (BoW).
- Initially used a Language Model for Labelling (LLM), but transitioned to a cost-efficient machine learning model using AWS Lambda and SageMaker due to budget constraints.
- Developed modules for sentiment, emotion, intent prediction, and customer vulnerability prediction based on the BoW model.
- Stored data in AWS S3 in parquet format and visualized insights daily via PowerBI.
- Integrated the solution with Git for version control and smooth productionalization.
- Delivered the solution within budget, achieving over 72% accuracy and over 60% coverage (for intent predictions with a probability threshold of more than 0.8).
Lead Data Scientist
Infosys Ltd.
08.2022 - 09.2023
- Led a full-cycle fraud detection project on Azure MLOps, from data acquisition to cleaning, and exploratory data analysis.
- Trained a variety of models including LightGBM, Random Forest, CatBoost, XGBoost, AdaBoost, and Azure AutoML.
- Built an automated retraining and inferencing pipeline using Azure MLOps, with a system for registering the best model based on a challenger vs. champion approach.
- Managed data efficiently by ensuring all results were automatically stored in Azure Blob Storage.
- Presented final predicted results via PowerBI, with monthly updates for the most accurate insights.
- Achieved substantial success in fraud detection, identifying 86% of fraudulent activity within the top 10% of high-risk customers, and 45% within the top 1%.
- Extended the model’s use case for real-time loan approval and credit scoring, and batch inference for overall fraud monitoring.
- Conducted comprehensive checks for model bias with respect to age, gender, and nationality, using FPR and TPR as key metrics, and passed all governance and compliance checks.
Lead Data Scientist
Infosys Ltd.
08.2021 - 06.2022
- Developed a proof-of-concept (PoC) for a cost of living crisis, creating a monthly aggregated database of customer spending, transactions, balances, and profiles using Teradata SQL.
- Identified customers in financial crisis based on overdraft charges and collection periods.
- Utilized three months of transactional and profile data to build a Balanced Random Forest classification model.
- Successfully identified over 70% of customers in financial crisis within the top 10% of high-risk customers.
- Leveraged Teradata Jupyter notebooks for model development, demonstrating proficiency in advanced data science tools.
Education
MBA - Finance And Analytics
NMIMS
Bengaluru, India04.2021
Bachelor of Science - Electrical Engineering
LDRP ITR
Gandhinagar, India06.2015
Skills
- Python
- Teradata - SQL
- PowerBI
- MLFlow
- Azure MLOps
- AWS MLOps
- Machine Learning
- NLP
- GenAI LLM (GPT, Hugging Face, Open Source)
- Conversational RAG
- Streamlit
- Git/GitHub
Certification
- AI-102: Microsoft Certified: Azure AI Engineer Associate
- DP-100: Microsoft Certified: Azure Data Scientist Associate
- AZ-900: Microsoft Certified: Azure Fundamentals
- AI-100: Microsoft Certified: Azure AI Engineer Associate
- AWS ML Specialty: AWS Certified Machine Learning - Specialty
- AWS Cloud Practitioner: AWS Certified Cloud Practitioner
Languages
English
Hindi
Gujarati
Timeline
Lead Data Scientist
Infosys Ltd.
03.2024 - Current
Lead Data Scientist
Infosys Ltd.
03.2024 - Current
Lead Data Scientist
Infosys Ltd.
03.2024 - Current
Lead Data Scientist
Infosys Ltd.
09.2023 - 03.2024
Lead Data Scientist
Infosys Ltd.
08.2022 - 09.2023
Lead Data Scientist
Infosys Ltd.
08.2021 - 06.2022
MBA - Finance And Analytics
NMIMS
Bachelor of Science - Electrical Engineering
LDRP ITR
Similar Profiles
Ujjwal Apurva SinghUjjwal Apurva Singh
Service Delivery Manager at Infosys Ltd.Service Delivery Manager at Infosys Ltd.
Vadivukkarasi PVadivukkarasi P
Senior Consultant at Infosys Ltd.Senior Consultant at Infosys Ltd.
Anuj Kumar SrivastavaAnuj Kumar Srivastava
Senior Analyst - (Data Science) at Infosys Ltd.Senior Analyst - (Data Science) at Infosys Ltd.