Ayush Pramod kumar

BI Engineer | Data Engineer | Analyst | Consultant

About

I am a seasoned Business Intelligence Engineer with extensive experience in data warehousing, ETL development, and scalable data modeling. My expertise lies in designing and launching robust data pipelines, developing actionable metrics, and creating dynamic visualizations to drive operational excellence. At Amazon, I spearheaded initiatives that reduced delivery delays by 70% and cut data pipeline runtimes by 60%, showcasing my ability to optimize processes at scale. I excel in cross-functional collaboration, working with engineers, product managers, and data scientists to develop innovative solutions that align with business goals while ensuring data privacy, security, and compliance. My technical proficiency in Python, SQL, AWS, and advanced data visualization tools enables me to deliver impactful insights and streamline decision-making processes across teams.

Interested? Get in touch!

Key Skills

Data Warehousing & ETL Development: Expertise in designing, building, and launching data pipelines using tools like SQL, Python, and AWS services.
Data Visualization: Proficient in creating dashboards with Tableau, Power BI, and internal tools to present key insights.
Cross-functional Collaboration: Strong experience in working with engineers, product managers, and data scientists to define and execute data strategies.
Metrics Development: Skilled in designing KPIs and data models to monitor performance, optimize processes, and improve SLAs.
Data Privacy & Compliance: Adept at implementing controls to ensure security and compliance with data governance standards.

Technical Skills

Education and Experience

Designed and implemented scalable data models and visualizations for monitoring middle and last-mile operations, reducing delivery delays by 70%
Developed and maintained ETL pipelines to automate data flows, improving efficiency and supporting optimization of delivery processes
Collaborated with cross-functional teams to gather requirements, ensuring accurate metrics and actionable insights for operational improvement
Built and launched data pipelines from scratch to identify inefficiencies, streamlining delivery performance
Partnered with stakeholders to define Service Level Agreements (SLAs) for critical datasets, ensuring data quality and availability
Created utilities using Python to optimize ETL workflows, mentoring peers on efficient query design and data processing

Designed metrics and dashboards to identify and mitigate buyer abuse, reducing policy violations and concessions by 12%
Engineered a data excellence framework, decreasing pipeline runtime by 60% while enhancing reliability and scalability
Independently developed data extraction and transformation processes to support key initiatives, ensuring compliance with privacy and security protocols.
Conducted weekly business reviews, translating complex data into actionable insights for leadership
Partnered with engineers and product teams to design solutions for detecting identity fraud, leveraging advanced analytics for policy enforcement

Managed deployment of product at various client location from technical and logistic point of view.
Developed and maintained a program to extract data from UWB Devices and converted them into human readable format.
Helped various clients to save resources and capital amidst COVID by securing there workforce using Contact Tracing solution.
Developed visualization to make Contact Tracing report human readable.
Created scripts to automate the processes for Deployment team saving the hours of manual work

Developed the Job-IQ Portal for Syracuse University’s career services office, enabling the student body to make informed decisions for optimized job search. The portal was meant to customize job search and interview questions based on several criteria. It used ML to guide the search and served as a tool for interview preparation as well. We had a student user base of 25K+ students across all disciplines and degrees.
Built web scrapers in Python to streamline the data collection from several sources for supporting business needs
Devised procedures to source data from APIs into excel and automated data processing and transformation using Python
Created ETL data pipelines in Python to integrate data from cloud APIs, XML/Excel, SQL server to perform insert/update/delete operations to the NoSQL database after data processing
Performed in agile methodology, interacted directly with entire team provided/took feedback on design, Suggested/implemented optimal solutions, and tailored application to meet business requirement and followed Standards
Influenced the technology selection process and lead the charge for usage of NoSQL databases, in place of traditional relational SQL, to optimize the processes and resources.

Completed my Master's Degree in Information Management with Specialization in Data Science
Also completed Certification Of Advance Studies (C.A.S) in Data Science.
My time at Syracuse University gave me an opportunity to work as a Student Advisor for MSAB( Master's Student Advisory Board ) where I served as a peer-facilitator peer-facilitator for new graduate students, arranged campus tours and ensured smooth transition into Syracuse Community. I was also a point of contact for incoming international graduates for local campus grievances. Also, collaborated with Admissions Office for induction of incoming international graduate students

Developed the Model Eye, one stop application to help banks make the transition from using disparate tools to grown models to a single platform. It’s automated workflow and governance tools, coupled with its collaboration and automated alert mechanism makes it’s the preferred model risk management platform. We had a user-base of 8K across all the banking clients of the company.
Developed ModelEye to meet the banking industry’s needs for a robust model governance platform
Participated in full development life-cycle of ModelEye by maintaining code and documentation
Collaborated with Business Analysts to formulate KPIs for the product dashboard to help user understand customer's behavior
Engaged with stakeholders to validate the scope of platform and procure feedback

Mentored a team of 7 individual towards development and delivery of product to various clients
Supervised the web-application development process across multiple projects
Procured investments from local vendors in exchange for services provided by company to overcome product development cost
Coordinated implementation of product at client’s end by supplying engineers and initial setup at reduced cost
Liaised between local ISP and clients to provide nominal rates for clients; led to increase in client-base for company

Completed my Bachelor's degree in Information Technology with Major in Software Development.
Hosted various conferences in the university in capacity of Conference Chair.
Successfully managed various Cultural Festivals ( Rubaroo 2014 & Rubaroo 2015 ) hosted by SRM University as Chairperson and General Chairperson for the event.
Acted as Chairperson of Creative Department of the University for two years.
Designed and Developed various website for different events at the university.

Projects

Linkedin Scrapper

Created this program to help a friend for his sales calls. The program uses your linkedin credentials to scrape through the available email address using the google search which is not limited to but open to all kinds of searches. The programs creates excel file with the available data.
The program works in both headless and the other way to ensure the captcha exception in login
Try it out from github and if it helps, leave a comment or review.

• See GitHub Repo

FudgeMart Inc. Dashboard

Business Intelligence Solution for Online Retail and VOD company. Developed High Level and Detail Level Dimensional Modeling of a Data Warehouse for a merger of 2 companies. Developed a ROLAP database to be used as the Data Warehouse and MOLAP cubes with time dimension using SSAS. Developed SSIS packages in MS Visual Studio using SSDT to automate the ETL process for over 500,000 rows of data. Optimized Sales and Order Fulfillment dashboards to be updated every 15 minutes, improving performance by 75%

Corona Virus Dashboard

The recent increase in Corona Virus cases lead to research more about its data. I came up with this Dashboard which is purely coded in R as a hobby project and to let the world know the latest statistics about corona virus. The data is pulled from the John Hopkins University Research Data which is available on Github, this data is updated everyday by sources which includes WHO ( World Health Organization ). It took me almost a day to wrangle up this data to be understandable and readable by my dashboard application.

• See GitHub Repo
• Visit Dashboard

Intel Image Classification

This project was a part of IST 718 ( Big Data Analytics ) course and was presented as a final project towards the completion of the course. The dataset comprises of around 25000 images. The scope of the project was the classification of these images in 6 different classes namely forest, building, sea, streets, mountain and glacier. To first proceed with the project, image was resized into 150X150. SVM and CNN were trained for achieving the classification. Features were extracted using HOG ( Histogram of oriented gradient ) Feature Extraction which was then passed to SVM Classifier to achieve accuracy of around 77%. The CNN (Convolutional Neural Network ) was designed with around 13 layers and trained to achieve the prediction accuracy of 83%.

• See GitHub Repo

Generating Paraphrase Resource using News Tweets

We surveyed two existing paraphrase sources viz. DIRT and PPDB. It is inferred that the existing resources are not being updated regularly. We, therefore present an approach which guarantees that the resource will be constantly updated since we are querying news tweets on a daily basis to generate binary paraphrase pairs.
Due to regular update, the accuracy of our resource comes around 75%-82%. Our resource can be used for question answering, typing suggestions, and other NLP Tasks.

• See Project Report
• See the Project Poster

New York City Motor Vehicle Collision Analysis

The project was a part of Information Visualization course which included analysis of New York City Motor vehicle collisions using R Studio. The Analysis was done to answer the following questions, which are then represented through visualizations. The Visualizations were exported from R Studio and then edited in Adobe Illustrator to improve the look and feel.

The aim of this project was to inform the citizens of New York City about how to avoid accidents by avoiding the travel routes and causes which were prominent after analysis of the data.

• See Project Poster
• See GitHub Repo

Analysis Of United States Airline

The project focuses upon analysis and producing actionable insights for the airline in United States. The project is designed using R language and can be reused by passing the similar dataset of airline information. The aim of the project was to suggest the airline management about the changes which should be focussed upon to increase the customer satisfaction. Following are the few points which were performed for completion of the project.

• See Project Report

Real Estate Database Solution

The project aimed at easing the physical documents in the transaction of the property. The database developed was capable of recording transactions like Renting and Selling of the property by the property owner.I was the chief architect of the database which was developed in MSSQL. My duties included leading my project subordinates and deciding upon the attributes which should be taken into account for development of the database. I also wrote triggers and stored procedures for keeping the database redundancy free and synced.

• See Project Report

Home Automation using Alexa API and RF Recievers

Done as a hobby project, I was able to install amazon's Alexa api on Raspberry Pi which was then used as a voice assistant to handle electric fixtures at home. The project included working with RF transmitter and Receiver and working with GPIO Pins to transmit radio signals. The language used to code the GPIO was python and was designed in a way to handle exceptions by the means of Alexa Feedback Channel. The project included installation of relays in traditional electric fixtures to enable them to control the switches remotely.

Contact

Bellevue, Washington[email protected]

Scan to Call Me!

Scan to download my Contact Information