Aashish Joshi
EXPERTISE
BIO
I am a Data Science professional with a cross-disciplinary expertise in Machine Learning, Data Analytics and Agile Product Management. I have over 5+ years of progressive work experience serving both small-medium sized as well as fortune 500 companies across United States. With my comprehensive understanding and hands-on experience across the entire data science product lifecycle, I have helped companies design, develop and implement end-to-end data science and machine learning solutions by playing the role of both an individual contributor as well as a strategic partner. I graduated with Master’s degree in Engineering from the USF with a specialization in Data and Statistics in 2018. I am immensely passionate about aligning Data Science methodologies to drive business outcomes.
WORK EXPERIENCE
Senior Data Scientist / Technical Product Owner
X-Synapse
May 2024 - Present
Created a decision intelligence framework to expose ML model functionality. Implemented online user interface (WSGI) to deliver realtime insights to decision makers/non-technical stakeholders (with 98% reproducibility factor). Operationalized models for deployment in production by developing microservices (API) with custom end points tailored to users’ preferences. Encapsulated ML applications for portability and reproducibility via Dockers. Leveraged enterprise grade MLOps tools/frameworks to standardize data science workflows for scalability and efficiency. Implemented experiment tracking, model registry and version control to optimize resource utilization and model performance. Served as a strategic partner to bridge the gap between data science solutions and business to maximize impact. Created a product roadmap to drive initiatives across Marketing, Sales, Customer Success and Operations departments.
Senior Technical Product Owner – Data Science
Honeywell
October 2020 - May 2023
Created an organization-wide SDLC framework to systemize the data science/ML product development, implemented Agile Product Management framework to streamline execution (leading to accelerated delivery cycles, cutting technical overhead by 45%). Served as SME and development lead, transforming business requirements into full-fledged data science and ML service offerings. Led a team of 3 data scientists to transition projects from research to production environment (achieving a 100% project success rate). Facilitated and led cross-functional collaboration (across data science, engineering, management, infrastructure and UI/UX teams) to realize the big picture product vision via user stories, product backlog, deliverables and project milestones. Demonstrated visionary leadership by actively engaging with key stakeholders to iteratively enhance product features (by establishing close alignment between data solutions and organizational/client objectives).
Data Science Analyst
Acadia (formerly Lift361)
February 2019 - February 2020
Developed supervised, semi-supervised and unsupervised machine learning (ML) models to maximize ROI on strategic initiatives (leading to 34% increase in annual revenue) by optimizing and measuring model performance using various metrics. Created custom data science workflows to conduct data pre-processing and automate the selection of best performing model (yielding an 18% reduction in customer churn and 23% uptick in campaign engagement rates) across a host of data driven use cases. Orchestrated cross platform data workflows and designed real-time dashboards, translating technical analysis into actionable recommendations for senior leadership (reducing time to deliver insights by 30% and improve resource allocation). Conducted Exploratory Data Analysis (EDA) to support informed decision making via descriptive/prescriptive statistics, detecting outliers, handling missing data, improving data quality and identifying patterns/relationships within the data.
Data Scientist Associate - Intern
GLOBAL IT EXPERTS
February 2018 - February 2019
Developed computer vision models with Convolution Neural Network (CNN) architecture using TensorFlow, achieving a classification accuracy of 76% in identifying defective solar panels. Standardized research environment workflows in Jupyter notebooks (data gathering, pre-processing, feature engineering, exploratory data analysis/EDA, PCA, hyperparameter tuning, k-fold cross validation, data cleaning etc.) to support ML model development. Improved model performance by addressing data imbalances via advanced data pre-processing techniques including stratified sampling, image augmentation, flattening, pooling etc.
Data Scientist/Analyst
BTP Infoservice
September 2014 - June 2016
Developed and implemented a logistic regression machine learning model using Python libraries (Pandas, NumPy, seaborn, SciPy, Matplotlib, Scikit-learn). Achieved accurate customer churn prediction and identified high-risk customers, resulting in a 15% improvement in retention through targeted sales campaigns. Conducted end-to-end data analyses in an agile environment, including data gathering, precise requirements specification, and data manipulation. Collaborated closely with business executives to ensure alignment between data science initiatives and business goals. Utilized advanced visual analytics techniques to extract actionable insights from complex datasets. Identified key customers and market segments to drive strategic initiatives for improved retention and informed decision-making. Consolidated data-driven findings into impactful PowerPoint presentations, effectively communicating results to diverse stakeholders and client executives. Employed storytelling techniques to convey the value of data analysis and drive business growth. Engaged in ongoing collaboration within an agile environment, interacting with business executives to align data science initiatives with evolving business needs. Delivered valuable insights and actionable recommendations in a timely manner.
EDUCATION
Masters
University of South Florida, May 2018
Electrical Engineering
ELECTRICAL & ELECTRONICS ENGINEERING
Visvesvaraya Technological University, June 2016
Bachelors
CERTIFICATIONS
Machine Learning A-Z in Python
Udemy
Issued: 7/27/2018
Credential ID: UC-V4NL37ER
Data Preprocessing & EDA with Python
Udemy
Issued: 4/2/2025
Credential ID: UC-4e61d078-f75a-4abd-ae02-269e53d909a6
Unsupervised Machine Learning with Python
Udemy
Issued: 4/2/2025
Credential ID: UC-cd338949-ebd2-4cd4-b772-2a03e391b3c6
SAFe Agile Product Management
Scaled Agile
Issued: 10/27/2021
Credential ID: 80129927-2232
PORTFOLIO PROJECTS
Advanced Regression - California Real Estate Pricing
House prices tend to fluctuate every year, so there is a need for a system to predict house prices in the future. House price prediction can help the developer determine the selling price of a house and can help the customer to arrange the right time to purchase a house. Finalizing scope: We intend to predict the median monetary value of house prices in the districts of California based on leveraging the data science pipeline mentioned above. Data requirements: We will leverage the California Housing dataset obtained from the StatLib repository.
View ProjectAdvanced Classification - IBM HR Analytics
The ‘Administration’ department is responsible for majority of the operational expenses within an organization. Employees make up the bulk of the administration costs and the cost of hiring new employees is significantly higher than the cost of retaining existing ones. The HR department is tasked with identifying employees that are most likely to exit the organization in the near future. This will ultimately help the organization to devise retention strategies in order to improve brand loyalty and minimize operational costs. Hence, the goal is to predict employees that are most likely to churn by utilizing supervised machine learning technique(s). Finalizing scope: To predict the employees that are the highest risk of leaving the organization by leveraging the data science pipeline mentioned above. Data requirements: We will utilize the IBM HR analytics dataset obtained from Kaggle in it’s entirety.
View ProjectSentiment Analysis - IMDB Movie Reviews
Sentiment analysis the process of classifying vast amounts of text data into several classes depending on the underlying tone of the text. These classes can be (but are not limited to) either positive, negative or neutral. Business across various industries rely on sentiment analysis to derive actionable insights about their customers and devise strategies to ultimately increase revenue. This data mining application is based on natural language processing and hence, can be extrapolated across a variety of real-world use cases involving textual data such as spam email detection, analyzing customer reviews to devise retention strategies, conducting market research to improve customer satisfaction, brand management, data driven investment decisions for hedge funds etc. Finalizing scope: To create a multipurpose text classification system (Machine Learning model) that can classify text into various classes. Data requirements: We will utilize the complete IMDB movie reviews dataset obtained from Kaggle to predict whether a movie review was positive or negative.
View ProjectSklearn Pipeline Demonstration - Iris Dataset Analysis
Most machine learning models require us to perform several data preprocessing steps such as feature scaling, PCA etc. as seperate lines of code before actually training the model on the data. This approach is not necessarily the best when it comes to code readability. Machine learning pipelines can be very handy in this case as they enable a user to write and organize cleaner looking code and help with the ease of understanding. Also, when new data is introduced, the pipeline automatically performs all the necessary pre processing steps before training the model on the new data for further prediction - thereby iterating and automating all the processing steps. We will be using the iris flower dataset from the sklearn library in order to demonstrate a machine learning pipeline in this notebook and predict the species of the flower.
View ProjectBreast Cancer Analysis - PCA and Hyperparameter Tuning
Breast cancer is one of the most commonly occurring cancers worldwide and according to global statistics, it represents a very high percentage of cancer-related deaths. In this repository, our goal is to build a machine learning model that can differentiate benign breast cancer lumps from the malignant ones. We will use several ML algorithms and evaluate their performance and predictive power after applying PCA and hyperparameter tuning techniques in order to come up with a 'best fit' model trained on data collected.
View ProjectFlower Recognition - CNN Image Classification
In this repository, our goal is to build a Convolution Neural Network to classify the type of flowers from a total of 5 different categories.
View ProjectVerified Machine Learning Engineer
3-5 years of experience
Preferred commitment: Full Time
Languages
Skills
Impressed by Aashish Joshi's profile?
Take the next step and bring this top talent to your team
Hire Aashish for your team