Sudharshan Sunder Gowrisankar
Chennai
Summary
Seasoned Data Scientist with a proven track record at Blackstraw.AI, leveraging Python and Machine Learning to pioneer AI-driven solutions, achieving an 80% accuracy in invoice data extraction. Demonstrated leadership in optimizing deep learning models, showcasing exceptional analytical skills and a commitment to innovation in data processing and analysis.
Overview
5
5
years of professional experience
Work History
Data Scientist
Blackstraw.AI
Chennai
05.2024 - Current
- Developed AI-driven solutions for automating invoice data extraction, leveraging LLM fine-tuning, HTML-to-text conversion, and advanced scraping techniques to achieve 80% accuracy in parsing diverse email formats into structured JSON for analytics.
- Built an AI-based email parsing system using APIs, OCR models, and rule-based engines, automating email-to-JSON conversion, attachment processing, and error handling for efficient data management and reporting.
- Designed a vector-based system to extract and validate UPC IDs using mobile-captured images, implementing custom computer vision models (ResNet-15) to improve retrieval accuracy and reduce memory usage, achieving 70% MAP and 80% Top-K retrieval rates.
- Created a robust pipeline for processing large-scale email invoice data, utilizing Python, Pandas, Beautiful Soup, and PyTorch for scalable and efficient automation across diverse datasets.
- Led data augmentation and fine-tuning initiatives for deep learning models (Mistral 8B, Llama 3.1, Qwen 7B), optimizing hyperparameters and model performance for complex enterprise data extraction tasks.
Software Engineer - Data Science
Expleo
04.2022 - 05.2024
- Developed and Deployed HR Policy Chatbot Using AI Foundational Technologies: Built a Retrieval-Augmented Generation (RAG) chatbot leveraging LLMs (GPT-3.5), LangChain, and vector databases to provide accurate, contextually relevant answers to HR-related queries (e.g., leave, insurance). Processed HR policy documents (PDFs) by chunking, vectorizing, and storing them in a vector database for efficient retrieval. Integrated the solution into a Streamlit/Chainlit interface with real-time streaming for enhanced user interaction.
- Implemented RAG Pipeline with Advanced Prompt Engineering: Designed a RAG pipeline to retrieve relevant document chunks using vector databases and feed them to GPT-3.5 for generating responses. Utilized prompt engineering techniques to optimize output quality and evaluated pipeline performance using RAGAS metrics, ensuring high context relevance and accuracy.
- Developed SQL Schema Comparator Tool with Advanced SQL/Python: Created a Python-based web tool for comparing schemas and detecting anomalies (e.g., missing data, duplicates, mismatches) during SQL database migrations. Leveraged Advanced SQL and Python to process and analyze complex data models, ensuring data integrity. Generated detailed reports with visualizations and exported results in Excel for validation.
- Built Synthetic Data Generation Tool Using AI and Data Modeling: Designed a Streamlit-based web tool to generate synthetic data from CSV, TSV, SQL tables, and other formats. Utilized the Synthetic Data Vault (SDV) framework and Advanced Python to replicate patterns in original data, enabling realistic data generation for testing, training, and analysis. Enabled easy download or direct database upload for seamless integration.
- Leveraged AI Tools for Intelligent Automation and Data Analysis: Demonstrated a strong ability to implement intelligent assistants and automation using LLMs, LangChain, and RAG. Processed and analyzed complex data models using Advanced SQL/Python and other programming tools to deliver scalable, customizable, and user-friendly solutions.
- Enhanced Data Accessibility with AI-Driven Solutions: Focused on improving user experience by integrating AI foundational technologies like LLMs, vector databases, and prompt engineering into tools. Provided real-time interaction, detailed visual reports, and downloadable outputs, ensuring accessibility and usability across all developed solutions.
Data Analyst
Moving Walls
Chennai
10.2020 - 03.2022
- Clustering unknown latitude - longitudes with known latitude - longitudes to separate groups.
- Segment and demography analysis of consumers.
- CRUD Operations of Summary Information.
- Campaign reports and data visualization.
- Creating automated reports on platform.
- Generating insights out of the reports with regards to audience movement data and their profile.
- Creating Automated Scripts in Python linking with Redshift and MongoDB
- Research Study projects
Data Science Research Intern
Tech Mahindra
Hyderabad
02.2020 - 08.2020
Education
Post Graduate Program - Data Science And Engineering
Great Lakes Institute of Management
Chennai12-2019
Bachelor of Engineering - Electrical And Electronics Engineering
Panimalar Engineering College
Chennai05-2018
Master of Technology (pursuing) - Artificial Intelligence
SRM Institute of Technology
ChennaiSkills
- Python
- SQL
- Machine Learning
- Natural Language Processing
- Deep Learning
- Generative AI
Accomplishments
WOW Award for Q2 2022 Best Performance
Timeline
Data Scientist
Blackstraw.AI
05.2024 - Current
Software Engineer - Data Science
Expleo
04.2022 - 05.2024
Data Analyst
Moving Walls
10.2020 - 03.2022
Data Science Research Intern
Tech Mahindra
02.2020 - 08.2020
Post Graduate Program - Data Science And Engineering
Great Lakes Institute of Management
Bachelor of Engineering - Electrical And Electronics Engineering
Panimalar Engineering College
Master of Technology (pursuing) - Artificial Intelligence
SRM Institute of Technology
Similar Profiles
VENKATESH BVENKATESH B
Scrum Master at STRAIVEScrum Master at STRAIVE
Barath NachimuthuBarath Nachimuthu
Aviation Officer at Bharat Stars Services Private LimitedAviation Officer at Bharat Stars Services Private Limited
Ranjani KasinathanRanjani Kasinathan
Assistant Vice President at CitiAssistant Vice President at Citi