Kedar Kanhere

• Leading a team of 4 data scientists while actively contributing to the hands-on development of ML and AI models.
• Designed and implemented a custom query-tuned chatbot for internal product use, enabling integration and querying across Wiki, SelectStar, and Jira, leveraging Anthropic’s Claude to enhance product understanding and efficiency.
• Developed an Optimal Take Rate model, boosting revenue by $2 million per quarter and reducing margins by approximately 2% using ensemble methods.
• Built an in-house A/B testing framework to evaluate model performance and developed an MLOps platform for Triton model integration, enabling the deployment of multiple models on GPUs, reducing latency, and achieving cost savings of $100K per month.

• Led other product and data science functions at ProjectPro, a Sequoia-funded startup, expanding the team from 3 to 47 employees.
• Created a custom fine-tuned LLaMA Index chat for answering patent and IP-related questions for the largest pharmaceutical company in the world.
• Developed a recipe and tutorial recommender system on the web with 438K+ users.
• Built the product from the ground up, increasing revenue to $300M and renewals by 60% over the period. Also worked on reforming, designing, and managing the product.

• Spearheaded the Data Science team, solving complex problems such as Multi-Objective Optimization and working on constraint optimization problems to determine the optimal revenue share for flights.
• Developed a generic data pipeline and in-house ML framework to enable seamless deployment of machine learning models.

• Worked as a Senior Data Scientist, tackling complex NLP problems related to OTC (Over the counter) documents.
• Built ML models and pipelines with high accuracy to classify OTC documents, which helped the organization save $500K in revenue each quarter.
• Developed a framework for other ML projects to simplify model deployment and facilitate insight analysis using MLFlow and Kubernetes

• Served as a Senior Data Scientist at WK, leveraging expertise in data analysis and statistical modeling to drive business growth and support informed decision-making.
• Developed the first-of-its-kind chatbot at WK, utilizing tabular data to answer customer queries, reducing wait times by 64% and resulting in a revenue gain of approximately $1.2 million. The chatbot was created using LSTM, Graph Networks (DataBase), and other technologies to provide responses in a conversational manner.
• Creating a time series forecasting model improving the prediction gain by ~10% in turn reducing the cost of personnel employed for support

• Created an in-house time series forecasting system deployed for weekly-level forecasts using models like ARIMA, ETS, and Prophet (Pepsi and Coke).
• Applied statistical modeling techniques to identify patterns and trends in data, enabling more accurate forecasting and improving decision-making (Walmart Forecast Picker)
• Developed and implemented machine learning algorithms to automate and optimize business processes and make cross connections (Twitter feedback for IKEA)
• Collaborated with stakeholders to understand their needs and delivered data-driven insights to drive business growth (Scotts Tools)

Here’s the corrected version:

• Started my career as a Trainee Decision Scientist at Mu Sigma Inc., gaining hands-on experience in data analysis and statistical modeling.
• Collaborated with cross-functional teams to develop and implement data-driven solutions.
• Utilized tools such as Python, R, and SQL to extract, clean, and analyze data.
• Assisted senior data scientists in analyzing and interpreting large datasets.
• Worked on various projects involving data analysis, statistical modeling, and machine learning algorithms.
• Collaborated with team members to analyze and interpret data, contributing to key insights and solutions.