Nived P P
Bengaluru
Summary
Data Science professional with 2+ years of experience in Generative AI development and Marketing Mix Modeling. Skilled in leveraging interdisciplinary knowledge to drive insights and bring value to work.
Overview
2
2
years of professional experience
Work History
Associate - Artificial intelligence and Analytics
Cognizant Technology solutions
Bengaluru
06.2023 - Current
- Company Overview: Information technology /Transactions
- Website: [Cognizant](https://www.cognizant.com)
- Developing code conversion tool using Generative AI in order to convert Azure synapse SQL into DBT SQL
- Iterative development of code base by understanding the client expectations
- Converting long and complex synapse SQL code into meaningful chunks by creating custom chunking strategies
- Appending helper parameters in code in order to make effective output debugging
- Create effective Prompt to convert the code into required format with 90%+ accuracy in conversion
- Testing the code at each point to verify the output consistency
- Developing a Generative AI-based root cause analysis chatbot for a Telecom client to reduce manual efforts in identifying fallouts in the data migration process
- Interact with US counterparts and understand the client requirements in Root cause analysis automation
- Utilized the Gemini API to create a chatbot, integrating Google GenAI embeddings to get context aware responses
- Employed stuff type conversational chain to manage and incorporate prompts and context within the chatbot's interactions
- Implemented FAISS (Facebook AI Similarity Search) for efficient vector store creation, enabling fast and scalable similarity searches
- Assisted in designing the user interface using streamlit, ensuring a user-friendly and visually appealing interface
- Conducted iterative fine-tuning of the chatbot model by analyzing performance results
- Assessed the model's accuracy through individual F1 scores, refining the model based on these metrics to improve overall performance
- Increased root cause identification efficiency by 100%, reducing manual effort from 60 to 120 root causes per hour, resulting in a 50% net gain in productivity
- Prepare and deliver detailed presentations to effectively communicate updates in model development and project progress to stakeholders
- Information technology /Transactions
- Website: [Cognizant](https://www.cognizant.com)
- Created a chatbot to find Capital expenditure details using financial reports for Fortune 500 companies using Retrieval Augmented Generation
- This is done using a PDF reader and chunking it with recursive character text splitter with a higher chunk size and medium overlap
- Vector store is created using FAISS and front end is modified with streamlit UI
- Actively engages in multiple outreach programs within the organization, promoting collaboration and knowledge sharing across teams
- Co-ordinate with Panelists, candidates and HR team to plan and organise lateral hiring for critical roles
Analyst
Ipsos MMA
Bengaluru
02.2023 - 06.2023
- Analyzed sales and marketing data to identify key performance drivers using marketing mix modelling
- Optimized marketing spend and improved ROI through data-driven insights like Rsquare and Mean absolute percentage error
- Collaborated with teams to align strategies with business goals
Education
Master of Arts - Actuarial Economics
Madras School of Economics
Chennai08-2022
Mechanical Engineering
APJ Abdul Kalam Technological University
Kerala09-2020
Skills
- Generative AI
- Prompt Engineering
- LLM finetuning
- Machine learning
- Regression
- Time series forecasting
- Marketing mix modeling
- Python
- PostgreSQL
- STATA
- PowerBI
- MS Excel
- Azure machine learning studio
Projects
Crude oil price forecasting using ARIMA model
- Applied Univariate time series analysis to forecast the crude oil price data using ARIMA model. Used Augmented Dickey-Fuller test to confirm stationarity and Akaike information criterion for diagnostic checks.
Flight arrival delay prediction using linear regression
- Used Azure machine learning studio to predict the arrival delay of flights in a particular Airport by understanding the key parameters using linear regression.
Estimation of growth rate per capita GDP
- Used linear regression to estimate the per capita GDP of Thailand using 15 years of past data using standard linear regression equation.
Extracurricular Activities
Volunteer, All India People Science Network Field Assistant, Sansad Adarsh Gram Yojna
Languages
- English
- Malayalam
Timeline
Associate - Artificial intelligence and Analytics
Cognizant Technology solutions
06.2023 - Current
Analyst
Ipsos MMA
02.2023 - 06.2023
Master of Arts - Actuarial Economics
Madras School of Economics
Mechanical Engineering
APJ Abdul Kalam Technological University