Vaidehi Pandey profile photo 1 on Tal

Vaidehi Pandey

Senior Data Scientist @ Qualys

Pune District

7+ years of work experience

Qualys
Jun 20262 mosPresent

Pune District, Maharashtra, India

Qualys
Apr 20242 yrs 3 mosJun 2026

Pune District, Maharashtra, India

— Natural Language Query to Qualys based tag language: • Finetuned open source models like mistral, llama using variations of LoRA. • Added guardrails to prevent users from asking irrelevant queries. • Evaluated the generated result using levenshtein distance and human evaluation. • Developed the inference script and wrapped it around an API to be consumed by the respective teams. — Catalogue Prediction • Developed a variation of hybrid RAG based pipeline to populate vendor and product for catalogue entries. • The developed system achieved accuracy of 80% on over 10,00000 entries. • Developed an API around this pipeline to be consumed by respective teams. — Qualys based chatbot • Developed a multi LLM based chatbot that converts the user query to elasticsearch DSL and in turn the DSL is used to retrieves the relevant data from the database. • Developed a retrieval system that extracts the relevant subset of index mappings based on the user query, which is then passed as context to the DSL generation pipeline. • Trained a bert based NER model to extract values and relationships from the user query and then create a prompt template for consistent results. • Implemented guardrails to prevent the chatbot from responding to out-of-domain questions or when data is unavailable in the specific DSL index. • Developed a RAG-based caching system that stores previous queries and their results, enabling faster retrieval of relevant information and enhancing the overall user experience. — Summarization • Developed a LLM based pipeline using few shot and COT prompting to find the affected vendor and product from the provided data. • Human evaluation was used as evaluation metric.
Resleeve
Sep 20221 yr 5 mosJan 2024

Eindhoven Area

Arspectra
Jan 20226 mosJun 2022

Luxembourg

Capgemini
Nov 20163 yrs 3 mosJan 2020

Pune, Maharashtra, India

collected , analysed and processed raw data scrapped from ticketing tool implemented a data pipeline to format data collaborated with support teams to automate business processes implement NLP techniques to classify tickets based on the description
Jan 20165 mosMay 2016

Greater Patna Area

studied various data formats of data as per requirement . Researched existing methodologies employed to develop data models and found scope of improvement. Thoroughly understood the Organisation structure of the Nationwide TB program and mapped the data to develop a data model

Vaidehi’s AI Stack

antigravity
antigravity
claude
claude

Education

University of Twente

Master of Science - MS, Data Science and Technology

University of Twente

20201 yr2021
sikkim manipal institute of technology

Bachelor of Technology - BTech, Computer Science

sikkim manipal institute of technology

20124 yrs2016

Join Vaidehi and 10k+ other professionals on

Tal Boss

Tal is the app where bosses meet engineers & builders directly, without middlemen

Download Tal Boss

Vaidehi Pandey on Tal

Headline

Senior Data Scientist @ Qualys

About

Experienced Machine Learning engineer with a proven track record in data-driven insights. Specialized in data science with a Master's degree. Proficient in machine learning and deep learning concepts.

Location

Pune District

AI stack

antigravity

claude

Work experience

Senior Data Scientist at Qualys, Jun 2026 - Present · 2 mos, Pune District, Maharashtra, India

Data Scientist at Qualys, Apr 2024 - Jun 2026 · 2 yrs 3 mos, Pune District, Maharashtra, India, — Natural Language Query to Qualys based tag language: • Finetuned open source models like mistral, llama using variations of LoRA. • Added guardrails to prevent users from asking irrelevant queries. • Evaluated the generated result using levenshtein distance and human evaluation. • Developed the inference script and wrapped it around an API to be consumed by the respective teams. — Catalogue Prediction • Developed a variation of hybrid RAG based pipeline to populate vendor and product for catalogue entries. • The developed system achieved accuracy of 80% on over 10,00000 entries. • Developed an API around this pipeline to be consumed by respective teams. — Qualys based chatbot • Developed a multi LLM based chatbot that converts the user query to elasticsearch DSL and in turn the DSL is used to retrieves the relevant data from the database. • Developed a retrieval system that extracts the relevant subset of index mappings based on the user query, which is then passed as context to the DSL generation pipeline. • Trained a bert based NER model to extract values and relationships from the user query and then create a prompt template for consistent results. • Implemented guardrails to prevent the chatbot from responding to out-of-domain questions or when data is unavailable in the specific DSL index. • Developed a RAG-based caching system that stores previous queries and their results, enabling faster retrieval of relevant information and enhancing the overall user experience. — Summarization • Developed a LLM based pipeline using few shot and COT prompting to find the affected vendor and product from the provided data. • Human evaluation was used as evaluation metric.

Machine Learning Engineer at Resleeve, Sep 2022 - Jan 2024 · 1 yr 5 mos, Eindhoven Area

Junior Machine Learning Engineer at Arspectra, Jan 2022 - Jun 2022 · 6 mos, Luxembourg

Associate Consultant at Capgemini, Nov 2016 - Jan 2020 · 3 yrs 3 mos, Pune, Maharashtra, India, collected , analysed and processed raw data scrapped from ticketing tool implemented a data pipeline to format data collaborated with support teams to automate business processes implement NLP techniques to classify tickets based on the description

Intern at National Institute of Technology, Patna, Jan 2016 - May 2016 · 5 mos, Greater Patna Area, studied various data formats of data as per requirement . Researched existing methodologies employed to develop data models and found scope of improvement. Thoroughly understood the Organisation structure of the Nationwide TB program and mapped the data to develop a data model

Education

Master of Science - MS, Data Science and Technology, University of Twente

Bachelor of Technology - BTech, Computer Science, sikkim manipal institute of technology