top of page
Professional Experience

KaggleX BIPOC Scholarship Program

  • Won the scholarship, which was given to only the top 20% of the participants.

  • Worked on designing models for analyzing key biomarkers of Alzheimer's Disease using MRI scans.

  • Gained comprehensive experience with GCP Vertex AI.

August, 2023 - November 2023

​

Utkarsh Lal KaggleX.jpg

StatusNeo

  • Working for Reliance Jio in the Artificial Intelligence Center of Excellence to develop predictive models for JioMart, Tira, and JioTV e-retail platforms. Mainly engaged in time series forecasting and clickstream data analytics.

  • Project 1: Product Popularity Forecaster for JioMart using SARIMAX: -
    Developed a SARIMAX model to forecast sales revenue and quantity of products on an e-retail platform called JioMart.
    Performed extensive feature engineering and tuned XGBoost for multivariate time series forecasting.
    Uncovered city-wise sales trends and seasonality.
    Created exogenous features to enhance forecasting.
    Stack - python, statsmodels, pmdarima, xgboost

  • Project 2: Product Popularity Ranker for Tira (e-retail platform): -
    Established the end-to-end Machine Learning pipeline for predicting product popularity of items on the Tirabeauty catalog.
    Tuned XGB and Huber Regressors on clickstream data to quantify product popularity based on user behavior and the relevance of click events like adds to cart and number of clicks.
    Implemented custom exponential smoothing with variable decay rates to induce logarithmic recency bias for data augmentation, favoring recent products.
    Created an Inferencing module to rank products based on popularity, thereby augmenting the efficacy of Auto-suggestions for search.
    Stack - Python, SQL, BigQuery, Scikit-learn, PyCaret

  • Project 3: Auto-Suggest Trending Content for JioTV: -
    Constructed an Autosuggest REST API for JioTV that recommends content based on trending searches and clicks in Cold Start scenarios.
    Reduced latency to 2ms.
    Built the PySpark ETL pipelines for processing data across Google Cloud Storage, Cassandra, and DataProc.
    Implemented a Geo-location module in Spark that translates IP addresses to City names using the MaxMind geolocation API.

    Ad-hoc project contributions: -
    Employed AWS Sagemaker and the BlazingText algorithm for sentiment classification of e-retail product reviews.
    Constructed ETL pipelines using PySpark in Azure Synapse.

August, 2022 - Jan 2024

Data Science Consultant

StatusNeo logo.png
Reliance JIO logo.png

American Express

  • Worked in the Credit and Fraud Risk Capabilities team for Amex’s Corporate card analytics. Engineered PySpark pipelines for Expected Credit Loss calculations.

  • Containerized and deployed pipelines using Docker and AWS Kubernetes.

  • Enhanced Credit Default and Fraud Detection Models using Random Forest and XGBoost by feature engineering, hyperparameter tuning, and Principal Component Analysis, decreasing run time by 50%.

Obtained the Google Cloud Professional Data Engineer Ceritification

March, 2022 - August, 2022

Analyst - Product Development

1200px-American_Express_logo_(2018).jpg
GCP DE badge.png

Accenture AI

  • Worked in the Applied Intelligence vertical of Accenture. Heavily involved in GCP - BigQuery, Vertex AI, DataProc, Data Fusion, DataFlow, Apache Airflow, Pub/Sub

  • Project 1: Customer Churn Predictor
    Developed a machine learning pipeline for Churn analysis and prediction for a telecom client.
    Created a streaming data pipeline handling 50,000+ rows per second of real-time data utilizing Spark Streaming.
    Employed Recursive Feature Elimination and Ensembling with Random Forest and XGBoost using soft voting classifier.
    Constructed batch Python data pipelines in Google Cloud Functions and DataProc. Orchestrated the pipelines using Apache Airflow and Cloud Composer. Performed extensive feature engineering using SQL in Google BigQuery.
    Interpreted the results and highlighted key features that indicated customer churn and visualized the insights through Data Studio dashboards.
    The insights drawn by interpreting the machine learning results directly informed significant business decisions, enhancing cell coverage and marketing strategies for customer retention.

  • Project 2: HealthCharge Classifier: Automated Expense Categorization for Healthcare Billing
    Tuned a LightGBM classifier for automated classification of healthcare expense categories.
    Built a corpus of over 100,000 keywords.
    Constructed a REST API as a quick inferencing module using TF-IDF vectorizer and cosine similarity index for text classification. Containerized using Docker and deployed on AWS EKS.

  • Adhoc Contributions
    Built pipelines in PySpark for Batch as well as Streaming data and employed Apache Airflow for scheduling and orchestration.
    Constructed web Crawlers using BeautifulSoup and Selenium for scraping product data from e-retail websites like Amazon.Worked in the Applied Intelligence vertical of Accenture. Heavily in

2560px-Accenture.jpg

August, 2021 - March, 2022

Data Engineering Analyst

Kimberly-Clark

  • Developed a Sentiment Analysis web-app for text mining and Sentiment Classification of online product reviews on Amazon and Shopee. Employed ML, Neural Networks, and multiple NLP concepts.

  • Reduced the net expenditure by over $60,000 per year by removing the need for outsourced services in the same domain.

  • Created and automated ETL pipelines in Alteryx and KNIME for Supply Chain Analytics.

  • Worked with clients from Russia, USA, and Argentina for automating their Supply Chain Analytics system

January, 2021 - July, 2021

Agile Analytics Engineer (Internship)

Kimberly-Clark_logo.jpg

Ernst & Young

  • Worked as a Software Engineer Intern to develop the web framework for the Ernst & Young Foundation Scholarship Program online tool.

July, 2020 - October, 2020

Summer Intern

1200px-EY_logo_2019.jpg

Amazon

  • Working as a Data Engineer in the Exports and Expansion BU.

March, 2024 - Present

Data Engineer

image.png
Volunteering Service

I worked at Prayas, which is a juvenile aid centre, where orphans and children in need are cared for. I taught music and science. I taught them about the construction and working of the different types of guitars (acoustic and electric). I showed them how an amplifier works and about the physics involved in it. Prayas, an NGO which is centred in Delhi is a haven for young children without a home. Founded in 1988, Prayas gave the homeless and troubled children of Jahangirpuri slums, shelter and a family. Today Prayas is spread all across India with multiple centres for juvenile aid.

​

bottom of page