KAUSTAV BHATTACHARJEE

Work History

Machine Learning Engineer

Tivian Kolkata

04.2022 - Current

Spearheaded the development of customized AI and ML solutions tailored to individual client needs, utilizing advanced generative AI and LLM technologies such as Llama, BERT and GPT
These solutions encompassed Semantic Topic Modeling, Fuzzy Text Matching, Sentiment Detection, Mood Detection, and other pertinent areas
Executed data-driven descriptive and predictive analytics to provide actionable insights for clients, ensuring informed decision-making and optimized strategies
Designed and implemented sustainable machine learning models focusing on feature importance, enhancing the efficiency and effectiveness of predictive analytics processes
Streamlined client data management through the utilization of Alteryx, automating workflows, and creating interactive macros for future scalability and efficiency gains
Developed project-independent applications within Alteryx for structured projects, maintaining workflows, and ensuring seamless project execution
Implemented predictive models utilizing machine learning tools such as Random Forest and Logistic Regression to assess issue severity and classify urgency, contributing to enhanced operational efficiency and risk management
Applied advanced techniques in text analytics using Spark NLP, deploying scalable models via StreamLit to extract meaningful insights from textual data
Conducted comprehensive analyses to identify and rectify data quality or integrity issues, as well as process or control gaps, ensuring data accuracy and reliability
Collaborated closely with clients to assess the current state of data quality within the designated domain, providing valuable insights and recommendations for improvement
Mentored and guided junior data analysts through shadowing and personalized feedback sessions, fostering professional growth and development within the team
Demonstrated effective communication skills in liaising with clients to understand their requirements, preprocessing data using Python, SQL, and Alteryx, and building and monitoring dashboards in Tableau
Optimized dashboard performance through query optimization and reconfiguration of table architecture in Tableau, ensuring streamlined data visualization and interpretation
Developed Tableau visualizations and dashboards from multiple data sources using data blending techniques, incorporating quick filters, parameters, and sets for enhanced usability and efficiency
Implemented user filters within Tableau workbooks to ensure controlled access, maintaining data confidentiality and security for sensitive information.

Manager

HSBC Kolkata

12.2020 - 04.2020

Spearheaded the meticulous documentation of the end-to-end project pipeline, meticulously capturing each phase from planning to implementation
This comprehensive documentation served as a guiding reference for team members, ensuring clarity and alignment throughout the project lifecycle
Implemented a systematic approach to prioritize deliverables, meticulously assessing their significance and impact on project objectives
This strategic prioritization optimized resource allocation, allowing the team to focus on high-impact tasks and maximizing overall productivity
Leveraged a deep understanding of team members' strengths, skills, and workload capacity to strategically distribute tasks
By matching responsibilities to individuals' capabilities, I fostered a collaborative and supportive team environment, enabling each member to contribute effectively to project success
Proactively identified and addressed potential bottlenecks and dependencies to expedite critical project components
This proactive approach minimized delays and risks, ensuring smooth project progression and timely delivery of key milestones
Employed effective project management techniques and proactive risk mitigation strategies to ensure timely project completion
By closely monitoring progress and adjusting strategies as needed, I consistently met or exceeded client expectations and organizational goals
Nurtured a culture of collaboration and continuous improvement within the team, fostering open communication and knowledge sharing
This collaborative environment facilitated effective task execution and goal attainment, driving overall project success
Maintained a sharp focus on delivering exceptional results within specified timelines, consistently exceeding client expectations and organizational objectives
This commitment to excellence and timely delivery has earned the trust and satisfaction of clients and stakeholders alike.

Business Intelligence Consultant

Tivian Kolkata

07.2019 - 11.2019

Managed user-level authentication processes to ensure secure access and data confidentiality within Tableau environments, implementing robust security measures to safeguard sensitive information
Utilized Tableau's flexibility to reconfigure table architecture as per specific dashboard requirements, optimizing data visualization and interpretation for enhanced decision- making
Demonstrated proficiency in building and publishing customized interactive reports and dashboards using Tableau Server, enabling stakeholders to access real-time insights and facilitating informed decision-making
Implemented advanced features such as action filters, parameters, and calculated sets in Tableau to prepare dynamic and interactive dashboards and worksheets, enhancing user engagement and data exploration capabilities
Enforced data security measures through the use of row-level security and user filters in Tableau, ensuring that sensitive data is restricted to authorized users only, thereby mitigating privacy risks
Developed Tableau visualizations and dashboards using Tableau Desktop, leveraging the platform's capabilities to create compelling data stories and actionable insights
Integrated multiple data sources seamlessly through data blending techniques, enabling comprehensive analysis and visualization of complex datasets within Tableau workbooks
Generated dashboards with quick filters, parameters, and sets to handle views more efficiently, improving user experience and facilitating rapid data exploration and analysis
Implemented context filters and data source filters to handle large volumes of data effectively, optimizing dashboard performance and responsiveness
Designed dashboards with measures including forecasts, trend lines, and reference lines, providing stakeholders with actionable insights into key performance metrics and trends
Implemented user filters within Tableau workbooks to control access and ensure data confidentiality, enabling appropriate teams to view and interact with relevant reports and dashboards securely.

Tableau Consultant

LAVA International Ltd

08.2014 - 03.2019

Played a pivotal role in database management, creating and optimizing database objects such as tables, views, procedures, triggers, and functions using T-SQL
These objects were designed to provide structure and definition to the data, ensuring efficient data maintenance and management
Utilized Tableau Server to build and publish customized interactive reports and dashboards, enabling stakeholders to access real-time insights and make data-driven decisions
Additionally, managed report scheduling to ensure the timely delivery of critical information
Implemented advanced features in Tableau, including action filters, parameters, and calculated sets, to prepare dynamic and interactive dashboards and worksheets, enhancing user engagement and data exploration capabilities
Implemented robust data security measures in Tableau, including row-level security and user filters, to restrict access to sensitive data and ensure compliance with privacy regulations
Developed Tableau visualizations and dashboards using Tableau Desktop, leveraging its intuitive interface and powerful capabilities to create compelling data stories and actionable insights
Integrated multiple data sources seamlessly through data blending techniques, enabling comprehensive analysis and visualization of complex datasets within Tableau workbooks
Provided 24/7 production support for Tableau users, ensuring uninterrupted access to critical business intelligence tools and resolving any technical issues promptly to minimize downtime
Created and deployed SSIS packages, utilizing various transformations such as Slowly Changing Dimension, Multicast, Merge Join, Lookup, Fuzzy Lookup, Conditional Split
Aggregate, Derived Column, and Data Conversion Transformations to streamline data integration processes
Optimized SQL queries for maximum efficiency and performance, fine-tuning database operations to improve overall system responsiveness and scalability
Designed and implemented a robust data warehouse architecture for corporate circulation Datawarehouse, establishing confirmed facts and dimensions to facilitate subsequent data marts, extracts, and reports
Developed Drill-down and Sub-Report functionalities using SSRS, monitoring report performance and ensuring timely delivery of actionable insights to stakeholders
Implemented advanced filtering techniques, including quick filters, parameters, and sets, to enhance dashboard usability and efficiency, enabling users to navigate and analyze data more effectively
Implemented context filters and data source filters to manage large volumes of data efficiently, optimizing dashboard performance and responsiveness
Designed dashboards with advanced features such as forecasts, trend lines, and reference lines, providing stakeholders with actionable insights into key performance metrics and trends
Published Tableau workbooks with user filters to control access and ensure data confidentiality, allowing only appropriate teams to view and interact with relevant reports and dashboards securely.

Application Developer

Sasken Communications

07.2014 - 08.2014

Symbian S40/S60, Windows 8 and Windows 8.1 and Open Source android (4.4.X) for East India.

Test Engineer

Marquistech Pvt.Ltd

02.2013 - 06.2014

Symbian S40/S60, Windows 8 and Windows 8.1 and Open Source android (4.4.X) for East India
Projects
Mood & Topic Detection Mar
2024, May 2024
The task entailed processing open-ended text data to discern both the sentiment and mood conveyed within the text
To accomplish this, I developed a robust solution employing two distinct models
Specifically, I engineered a Transformer pipeline function capable of seamlessly integrating with data frames
This function augmented the original data frame with three additional columns: Sentiment, Mood, and Mood_score
Employing the HuggingFace API, I leveraged pre-trained models tailored to the task's specifications
In addition, I implemented optimization techniques to enhance the efficiency of the pipeline
Notably, I utilized the "bitsandbytes" quantization pipeline to refine model efficiency, reducing resource consumption by quantizing the model to a 4-bit format
This optimization strategy effectively mitigated resource demands, ensuring optimal performance without compromising on accuracy or functionality.

Feature Importance

01.2023 - Current

In this project, our focus was on analyzing customer data to discern the most influential factors contributing to customer churn within the organization
To achieve this, we employed a comprehensive approach combining Variance Inflation Factor (VIF) analysis and Significance Value assessment
We utilized both Logistic Regression and Random Forest models to identify and prioritize the key features driving customer churn
Logistic Regression allowed us to assess the statistical significance of individual predictors, while Random Forest analysis provided insights into feature importance based on predictive power
By integrating these methodologies, we aimed to develop a robust understanding of the factors influencing customer attrition
This approach enabled us to identify and prioritize actionable insights, empowering the organization to implement targeted retention strategies and mitigate customer churn effectively.

Image Comparison

04.2024 - 05.2024

The objective was to develop an OpenCV-based Optical Character Recognition (OCR) model capable of comparing images based on their structural and informational attributes
The aim was to generate a probabilistic data frame indicating the similarity between images
To achieve this, I employed a combination of an OCR-based model and regular expressions to assess matches
The process began by reading images into CV2 and converting them into grayscale before inversion
Subsequently, Pytesseract was utilized to extract text from the images, with a threshold applied to evaluate the significance of the match
This approach facilitated the creation of an efficient and accurate OCR model capable of comparing images and generating probabilistic data frames to quantify their similarity.

Text Analysis

04.2023 - 12.2023

Within this project, I collaborate closely with a data scientist from our team to develop macros utilizing advanced text analytics models
Our primary objective is to engineer scalable and interactive modules capable of seamlessly integrating into various projects as per client specifications
The overarching goal is to design robust solutions that enable the extraction of valuable insights from open-ended textual data
These modules are meticulously crafted to cater to the diverse needs of our clients, empowering them to derive actionable intelligence from unstructured text data
Through this collaborative effort, we aim to enhance workflow efficiency and facilitate informed decision-making by leveraging cutting-edge text analytics techniques
Our focus is on creating versatile tools that streamline the process of extracting meaningful information from textual data sources, thereby adding significant value to our projects and ultimately, to our clients' businesses.

Smapshot Generator

Tivian

11.2022 - 03.2023

In this project, our objective was to address the challenge of prolonged processing times and increased loading durations when aggregating or slicing data in Tableau, particularly concerning larger datasets
To achieve this, we embarked on a mission to create snapshots encompassing all possible combinations of existing demographics
These snapshots serve as comprehensive reference points, enabling efficient slicing and dicing of data
By pre-computing these combinations, we effectively reduce the computational overhead during data analysis in Tableau, leading to significant improvements in processing times
Our goal was to streamline the data exploration process by providing users with optimized datasets tailored to their analytical needs
This approach not only enhances the overall performance of Tableau dashboards but also empowers users to derive insights more rapidly and effectively
Through this project, we aimed to enhance the usability and responsiveness of Tableau dashboards, ultimately facilitating more efficient decision-making processes within our organization.

ENBW

03.2023 - 02.2024

In this project, my responsibility entailed the collection and aggregation of data derived from customer surveys, a pivotal task in driving data-driven insights
Leveraging Alteryx, I meticulously gathered and consolidated survey data, ensuring its accuracy and completeness for further analysis
Additionally, I spearheaded the development of macros aimed at automating historical data input processes
One notable feature of these macros was the incorporation of a parameter mechanism, providing users with the flexibility to limit historical data directly at the data source level
This innovative approach not only streamlined the data extraction process but also significantly reduced data load times within Tableau dashboards
By implementing these automation solutions, I contributed to enhancing the efficiency and usability of our analytics workflows, ultimately enabling stakeholders to access and analyze critical insights with greater speed and precision
This initiative underscored my commitment to driving operational excellence and delivering actionable insights to support informed decision-making within the project
Data

South Central Ambulance

04.2022 - 12.2022

In this project, I spearheaded the development of a comprehensive end-to-end dashboard solution tailored to meet the client's needs
The dashboard provides a holistic view of key performance indicators (KPIs) relative to previous periods, utilizing a quarterly marker for comparison
Additionally, I incorporated a demographic split feature to highlight the favourable percentages across various demographic segments, enabling deeper insights into customer behaviour and preferences
To enhance data interpretation, I integrated a thematic filter for open-ended comments, organizing them based on relevant themes
This feature facilitates efficient analysis and enables stakeholders to identify prevailing trends and concerns
Recognizing the importance of performance optimization, I continually refined table queries to enhance dashboard responsiveness and speed
By leveraging Alteryx, I automated data processing and scheduling, ensuring seamless integration with
Tableau Server for real-time data updates and accessibility
Through these efforts, I delivered a robust and user-friendly dashboard solution that empowers stakeholders with actionable insights, facilitates informed decision-making, and drives organizational success.

Ontology, HSBC

12.2021 - 01.2022

The primary objective of this project was to develop a user-friendly front end capable of dynamically generating SQL queries based on user selections of database and table names from the Common Data Model (CDM)
By creating an intuitive interface, users could effortlessly navigate through available databases and tables within the CDM
The front end provided a seamless experience for users to select specific database and table combinations, allowing for the customization of SQL queries to extract relevant data
This solution aimed to streamline the data retrieval process, eliminating the need for manual query construction and reducing the likelihood of errors
By empowering users to generate SQL queries through a simple selection process, the front end facilitated efficient data extraction from the CDM, ultimately enhancing productivity and accelerating decision-making processes.

Summary

Overview

Work History

Machine Learning Engineer

Manager

Business Intelligence Consultant

Tableau Consultant

Application Developer

Test Engineer

Smapshot Generator

Education

PG Diploma - Data Science

B.Tech - Electonics and Communication

Higher Secondary - undefined

Madhyamik - undefined

Skills

Awards

Toolsandframeworks

Work Preference

Work Type

Work Location

Important To Me

Software

Languages

Work Availability

Interests

Websites

Timeline

Similar Profiles

Pehr JacobsonPehr Jacobson

Christian KampChristian Kamp

Sahabuddin KhanSahabuddin Khan

Shubham NarsariaShubham Narsaria

Vikrant KumarVikrant Kumar