Still Struggling with Tech Hiring? Discover Olibr's Solution Now!
Login / RegisterDeveloperClient
Login/Register

MALLIKARJUN CHAULAGI

Data Scientist Data & Analytics

  • Bangalore Urban, KA
  • 12 years
Notice Period: 15 Days

    Candidate's About

    Over a decade in IT, specializing in Data Science (6+ years) covering Machine Learning, Data Mining, and Statistical Analysis, combined with 8 years across Product Management, Development, Data Analytics, and Database/ETL roles.

    Work & Experience

    D

    Data Scientist Data AND Analytics

    Nov 21 - PRESENT

    Technology

    • Icon
      Tableau
    • Icon
      Azure
    • Icon
      SQL
    • Icon
      Python

    Tools

    • Icon
      Bootstrap
    • Understand data and the different platforms used by Bekaert.
    • Translate business requirements from the client into equivalent modeling tasks.
    • Actively manage and lead ongoing predictive modeling assignments.
    • Engage with clients, understand requirements, and suggest predictive modeling and machine learning solutions.
    • Build and improve Recommendation Engines, P...
    Read More

    Bekaert

    Data Scientist Data & Analytics

    • Lead predictive modeling projects translating client needs into tasks.
    • Enhance Recommendation Engines, Propensity Modeling, etc.
    • Manage data science team and collaborate with engineers.
    • Analyze large user data, establish cause & effect, and drive data-driven decisions.
    • Promote data science across the organization and educate peers.
    • Data tracking
    • Data Pipelines
    View Details
    T

    Technical Architect Data Scientist

    Jan 20 - Oct 21

    Technology

    • Icon
      SQL
    • Icon
      Azure
    • Icon
      Python
    • Icon
      Tableau

    Tools

    • Icon
      Bootstrap
    • Acted as a Technical Architect Data Scientist at HCL Technology, responsible for providing technical leadership and expertise in data science projects.
    • Led the design and architecture of data science solutions, ensuring alignment with business goals and requirements.
    • Collaborated with cross-functional teams to gather and analyze requirements, translating them into technical...
    Read More

    Shipment Property Anomaly Detection (SPAD)

    Technical- Architect (Data Scientist)

    • Implemented AI and ML techniques to minimize penalties and fines associated with incomplete shipment information.
    • Utilized AI algorithms to improve the accuracy and completeness of shipment information.
    • Employed ML models to automate the verification and validation of documents for customs clearance.
    • Developed predictive analytics models to identify potential issues in advance and take proactive measures.
    • Leveraged AI-powered systems to ensure correct duties and taxes are paid based on accurate information.
    • Integrated AI solutions to validate the International Harmonized System Code for shipments.
    • Analyzed historical data to identify patterns and trends related to penalties and fines, enabling preventive actions.
    • Collaborated with stakeholders to streamline processes and enhance data accuracy and completeness.
    • azure
    • sql
    View Details

    Document Identifier

    Technical- Architect (Data Scientist)

    • The multi-class model was used to predict the document type when uploading supporting documents for shipping an item.
    • The OpenCV OCR (Optical Character Recognition) technology was employed to extract text from the uploaded documents.
    • The extracted text was then fed into the XG Boost algorithm, which had been trained on a labeled dataset to classify documents into different types.
    • By analyzing the content of the documents, the model could suggest the appropriate document type to the customer.
    • This proactive suggestion aimed to prevent customers from mistakenly uploading documents as type 'OTHER,' which could increase the risk of shipment delays or additional expenses for the client.
    • The proactive suggestion feature helped streamline the document submission process and improve overall efficiency in document handling and shipping operations.
    • azure
    • sql
    View Details

    Named Entity Recognition (NER)

    Technical- Architect (Data Scientist)

    • The NER (Named Entity Recognition) model was implemented using SpaCy to identify common entities in shipment documents.
    • The model was trained on a labeled dataset to recognize entities such as names, locations, document names, product descriptions, and Harmonized System Codes (HS codes).
    • SpaCy's pre-trained language models, such as en_core_web_sm, were used as a starting point for training the NER model.
    • The training data consisted of annotated examples where entities of interest were manually labeled.
    • The NER model was fine-tuned using the labeled dataset to improve its ability to accurately identify specific entities in shipment documents.
    • During inference, the NER model was applied to the shipment documents to extract structured data by identifying and labeling entities.
    • The extracted entities provided valuable information for various tasks such as document classification, data analysis, and automated processing in the shipment workflow.
    • By automating the extraction of entities, the NER model reduced manual effort, improved data accuracy, and facilitated efficient information retrieval from shipment documents.
    • azure
    • sql
    • AzureSql
    View Details

    Cage Risk Mode

    Architect (Data Scientist)

    • RF Classifier (Binary): A predictive model built to reliably predict the risk of a package being caged during different stages of a shipment, with an emphasis on identifying the possible cause/s.
    • MLP Classifier (Multi Label): Another predictive model used to identify multiple causes associated with the risk of a package being caged during various shipment stages.
    • Azure
    • Sql
    View Details

    Cage Dwell Model

    Architect (Data Scientist)

    • XG Boost Model: Developed a machine learning model using XG Boost algorithm to predict the number of days an "at risk" shipment will be held in clearance during international shipping.
    • The model provides advance notification for proactive intervention, helps recalculate the Estimated Date of Delivery, and notifies both the client and FedEx operations for efficient management of the shipment process.


    • Azure
    • Sql
    View Details

    Restricted Party Service(RPS)

    Architect (Data Scientist)

    • RNN-LSTM Model: Implemented a deep learning model using RNN (Recurrent Neural Network) with LSTM (Long Short-Term Memory) architecture.
    • The model is used in the RPS (Restricted Person Screening) system to check various lists from governmental organizations and FedEx sources.
    • It identifies if a given entity/individual should be restricted from shipping commodities, thereby reducing FedEx's operational cost by 75%.


    • Azure
    • Sql
    View Details

    Entity Profiling

    Technical- Architect (Data Scientist)

    • Developed dedicated Feature Stores for Location, Commodity, and Party Profiling.
    • Transformed and aggregated historical shipping data into Feature Store tables.
    • Data utilized by Cage Risk Models for analysis.
    • Applied K-Means Clustering to generate additional features.
    • Segmented similar behavior in Location, Commodity, and Party categories.
    • K-means clustering
    View Details
    T

    Technical Architect Data Scientist

    Mar 19 - Jan 20

    Technology

    • Icon
      Azure
    • Icon
      SQL
    • Icon
      Python
    • Icon
      Tableau

    Tools

    • Icon
      Bootstrap
    • Migrated 250,000 clients of Mercedes Benz-MBRDI (Daimler AG) to Windows 10.
    • The Operations team handled 600-800 incidents per month.
    • Faced challenges such as manual resolution of recurring incidents, difficulty in determining incident severity, lack of standard solutions, and a lack of a streamlined knowledge repository.
    • Implemented analysis on the tickets and integ...
    Read More

    Logs

    Technical- Architect (Data Scientist)

    • Applied LDA (Latent Dirichlet Allocation) and tf-idf (term frequency-inverse document frequency) models to analyze incident logs.
    • Identified the top technology-based topics from the incident logs using these models.
    • LDA helped uncover latent topics by assigning probabilities to words, while tf-idf weighted words based on their importance in the documents.
    • By combining LDA and tf-idf, the model accurately extracted and ranked the most relevant technology topics from the incident logs.
    • This analysis provided valuable insights for understanding common technology-related issues and optimizing IT support processes.
    • azure
    • sql
    View Details

    Linear Classifier

    Technical- Architect (Data Scientist)

    • Developed machine learning models using SGD (Stochastic Gradient Descent) algorithm to determine the severity of incidents based on historical data.
    • Utilized historical incident data to train the SGD model, which is a popular optimization algorithm for training linear classifiers.
    • The trained model was capable of assigning severity labels to new incidents based on patterns and features observed in the historical data.
    • By accurately predicting incident severity, the models provided valuable insights for prioritizing and allocating resources effectively, leading to efficient incident resolution processes.
    • Azure
    • Sql
    View Details

    Management Efficiency

    Technical- Architect (Data Scientist)

    • Built a model for incident categorization using RNN (Recurrent Neural Network) with LSTM (Long Short-Term Memory) architecture.
    • Implemented the RNN-LSTM model to classify incidents into different categories based on their characteristics and patterns.
    • Trained the model using labeled incident data to learn the relationships and dependencies between incident descriptions and their corresponding categories.
    • The RNN-LSTM model enabled accurate and automated incident categorization, improving efficiency in incident management and resolution processes.
    • Azure
    • Sql
    View Details

    Routing System

    Technical- Architect (Data Scientist)

    • Developed an efficient routing system using the Random Forest model.
    • The Random Forest model was trained on relevant data to learn patterns and make informed routing decisions.
    • Leveraging the ensemble of decision trees, the model considered various factors such as traffic conditions, distance, and delivery priorities to optimize route planning.
    • The routing system successfully provided efficient and effective routes, reducing delivery time and improving overall operational efficiency.
    • Azure
    • Sql
    View Details

    Solution Recommender System

    Technical- Architect (Data Scientist)

    • Developed a solution recommender system using the cosine similarity model.
    • Implemented the cosine similarity algorithm to measure the similarity between different solutions and user needs.
    • The recommender system utilized this similarity measure to suggest potential solutions that closely matched the user's requirements.
    • By leveraging cosine similarity, the system provided accurate and relevant recommendations, enhancing the user experience and facilitating efficient problem-solving.
    • Sql
    • Azure
    View Details

    XGBoost

    Technical- Architect (Data Scientist)

    • Developed an incident volume forecasting model using XGBoost.
    • Trained the XGBoost model on historical incident volume data to learn patterns and trends.
    • Leveraged the powerful gradient boosting algorithm of XGBoost to accurately predict future incident volumes.
    • The model provided valuable insights for capacity planning, resource allocation, and overall incident management strategy.
    • Azure
    • Sql
    View Details

    Multinomial-Logistic Regression

    Technical- Architect (Data Scientist)

    • Developed a model using Multinomial Logistic Regression to analyze routing issues in ticketing systems.
    • The model aimed to determine the number of routing hops required for tickets and identify reasons behind tickets being routed from one team to another.
    • Trained the model using historical ticket data, considering various features such as ticket attributes, content, and previous routing patterns.
    • Multinomial Logistic Regression allowed for the prediction of routing hops and identification of factors leading to ticket re-routing, improving the overall efficiency of ticket handling and reducing delays.


    • azure
    • sql
    View Details

    Senior Consultant (Data Scientist)

    Jun 14 - Feb 19

    Technology

    • Icon
      Azure
    • Icon
      SQL
    • Icon
      Tableau
    • Icon
      Python

    Tools

    • Icon
      Bootstrap
    • Managed projects with a team of four data scientists and DB developers.
    • Extracted and validated data from marketing databases for analytical projects.
    • Assessed analytical projects to determine required data elements for desired results.
    • Collaborated with external database host vendors for project completion.
    • Developed Decision Tree and Deep Learning models f...
    Read More

    BAL & eServe Billing Application

    Senior Consultant (Data Scientist)

    • BAL & eServe Applications: Telecom Billing Application solutions.
    • Provides a centralized access point to the billing system.
    • Allows provisioning, modification, and termination of customers, accounts, and products.
    • Provides real-time access to up-to-date billing information.
    • Supports payment management, adjustments, invoice generation, and related billing operations.


    • Azure
    • Sql
    View Details
    S

    Senior software Engineer

    Oct 10 - Jun 14

    Technology

    • Icon
      Azure
    • Icon
      SQL
    • Icon
      Tableau
    • Icon
      Python

    Tools

    • Icon
      Bootstrap
    • Involved in System Requirement Specification (SRS) and Data element documents.
    • Developed PL/SQL packages and procedures for coupon payments, amortization, accruals, balance sheet, and maturity payments.
    • Developed NPT (Nostro Projection Tool) for balance amount aggregation in transactions.
    • Proficient in Data Modeling, Data Analysis, Data Warehousing, Data Mapping, En...
    Read More

    Market Bonds and Certificate of Deposits (CD’s)

    Senior software Engineer

    • BONDS module: Treasury Back Office function integrated with branch systems for STP benefits.
    • Handles fixed income bonds and Money Market instruments (CPs, CDs, Repos, Reverse Repos).
    • Supports deal life cycle from capture to processing and accounting movements.
    • Includes valuations based on regulatory requirements.
    • Handles products such as Central Government Bonds, Special Bonds, Certificate of Deposit, Repo, and Reverse Repo.
    • Azure
    • Sql
    View Details

    Education

    V

    B.E.

    Vishweshwaraiah Technological University
    2001 - 2005