About
I am a software developer with a background in Finance & Banking. My experience involves developing tools to assist data ingestion, financial planning, testing check processing systems. I have also done research involving social media and stock market movements. My hobbies include hiking and camping outdoors, surfing and scuba-diving, and enjoying art installations.
Software Developer & Finance Graduate.
- City: Henderson, Nevada
- Degree: Bachelors
- Email: mjacquot.dev@gmail.com
- LinkedIn: Marcel Jacquot
- Github: https://github.com/mjacquot1
Web Stack
This website is built using the following technologies.
Resume
Skills
Programming Languages
- RUBY
- VBA
- JAVASCRIPT
- PYTHON
- HACK (PHP)
Frameworks
- DJANGO
- RUBY ON RAILS
Infrastructure
- DOCKER
- NGINX
- GUINICORN
- GIT
- CELERY
- CLOUDFORMATION
- TERRAFORM
- AWS
- UNIDASH (TABLEAU)
- PRESTO SQL
- DAIQUERY (JUPYTER)
- DATASWARM (AIRFLOW)
- APACHE HIVE
Technologies
- RSPEC
- GRAPHQL
- REGEX
- PYTHON-DOCX
- XLWINGS
Databases
- MONGODB
- POSTGRESQL
PROFESSIONAL EXPERIENCE
SOFTWARE ENGINEER IC4, INFRASTRUCTURE- META
April, 2024 - February, 2025
Bellevue WA- Ran data analysis on trillions of rows in Apache Hive tables to discover and action violating harmful content.
- Designed an API to improve ML models for harmful content detection.
- API can handle approximately 100K daily searches and return tens of millions of results.
- Optimized a content management and review pipeline by 4%, saving 420 man-hours per year.
- Implemented a decision tree for classifying violations enabling an estimated 800K new content labels annually.
- Processed trillions of rows in Apache Hive tables to develop a heuristic with 98.5%+ accuracy using Daiquery (Jupyter fork) to identify and remove hashtags advertising illegal content.
- Led cross-functional efforts that resulted in the removal of over 126K policy-violating hashtags from Instagram and the identification of 600K+ posts for removal.
- Designed and deployed an internal Hack API to simulate Facebook Search to support ML engineers in proactively detecting harmful content, blocking search terms, and training models for improved violating search term detection.
- API can handle approximately 100K daily searches and return tens of millions of results.
- Enabled searching terms using both manual inputs and automatic generation via internal LLMs.
- Researched legacy code to ensure compatibility and shared findings to facilitate seamless integration.
- Designed and deployed a plugin to optimize a content management and review pipeline handling 300K jobs annually, improving efficiency by 4% and saving 420 man-hours per year.
- Implemented multiple pipeline steps in Dataswarm (Apache Airflow fork) using Presto SQL to analyze review history and eliminate redundant content reviews.
- Collaborated closely with the internal policy team to ensure compliance with review standards.
- Researched legacy code to ensure compatibility and shared findings to facilitate seamless integration.
- Implemented a decision tree for classifying violations related to high-risk content labels in content moderation, enabling human reviewers to assign newly unavailable content labels an estimated 800K times annually.
- Enhanced the accuracy of integrity ML models by routing appropriately labeled content to training pipelines.
- Collaborated with cross-functional policy and operations teams to ensure the solution met end-user needs.
- Enhanced content investigation workstreams by implementing unit tests, error handling for edge cases, and loggers with dashboards in Unidash (Tableau fork) to track performance and impact metrics.
- Authored research and design documents, updated internal programming guides and internal wikis to provide comprehensive documentation for new users of the developed features.
SOFTWARE ENGINEER L2, PLATFORM- DRIZLY / UBER
March, 2022 - March, 2023
Remote- Automated a manually intensive product attribute updating process using Ruby on Rails. Developed automated test scripts with RSpec.
- Developed and implemented both front-end and back-end features. Secured GraphQL queries and data ingstion.
- Configured and ran Docker images for local development. Employed infrastructure as code tooling such as CloudFormation and Terraform
- Automated a manually intensive product attribute updating process by developing and deploying an application feature that asynchronously updated product attributes from supplier-provided Excel workbooks.
- Developed and implemented both front-end and back-end features for updating product attributes utilizing Ruby on Rails gems including Views and Active Record.
- Designed and developed AWS Lambda Ruby functions. Employed infrastructure as code tooling such as CloudFormation and Terraform for VPC configurations, IAM security measures, and S3 connections.
- Secured GraphQL queries and mutations by implementing authorization hooks based on user credentials within a centralized code repository. Mapped inter-app GraphQL usage between multiple stacks.
- Strengthened code quality and reliability by developing and executing RSpec unit tests to validate program functionality, ensure accurate database updates, and effectively both handle and log errors.
- Monitored errors using New Relic alerts and implemented error handlers in critical areas where user input was accepted or database operations were performed. Generated error reports for internal users.
- Developed and implemented data scrubbing modules using Ruby on Rails to secure external data inputs and protect against Json, SQL, and VBA injection attacks.
- Configured and ran Docker images to facilitate local app development and testing.
- Created internal documentation outlining the configuration process for a Terragrunt/Terraform repository used to deploy Ruby Lambdas within the primary stack.
- Utilized Git for CI/CD. Conducted code reviews and approvals to ensure deployment readiness.
- Adhered to Agile methodologies throughout the software development lifecycle. Maintained open lines of communication with all project stakeholders and created project documentation using Jira.
LEAD SOFTWARE ENGINEER- LANDMARK WEALTH MANAGEMENT LLC
July, 2019 - March, 2022
Ontario, CA- Used python, VBA, and MongoDB to architect, design, and implement:
- A Python program to generate financial plans after automatically ingesing client information.
- Excel workbooks to automate day-to-day financial planning tasks. and calculations.
- A MongoDB database to hold and access client information.
- Developed an automatic document writer through Python UI and MongoDB integration, allowing client information to be centralized to programmatically generate personalized financial plans.
- Oversaw the continual development of an Excel and VBA macro based cash flow tool that provides tax calculations, projected client cash flows, and retirement account recommendations.
- Employed a NoSQL data model with MongoDB using the attribute design pattern.
- Identified design problems early to refactor the necessary application code ensuring long term extensibility.
- Architect, designed, implemented and deployed the following features:
- A Python 3 Desktop GUI application to generate formatted word documents based on a list of userdefined conditional arguments and formatting rules.
- An interface to allow users to easily create conditional logic, translating the string conditions to executable code.
- An interface to allow users to easily create formatting rules that translate into Microsoft Word syntax.
- A calculator interface to easily create formulas used for conditional logic.
- Template revision history and restore feature to allow users to revert to an older template.
- An interface utilizing a Breadth First Search algorithm to import an Excel workbook and parse user conditional variables defined in the template.
BUSINESS ANALYST- MITSUBISHI UNION FINANCIAL GROUP
December, 2017 - July, 2019
Monterey Park, CA- Managed development projects for item processing.
- Developed VBA Macros to assist in operations.
- Used Excel VBA macros to randomly generate values for Checking items used for testing, resulting in an increase in testing accuracy, faster defect resolution, and a 95% reduction in man-hours.
- Designed, developed, and tested the Canadian item processing capture platform for our location in Tempe, Arizona, and oversaw record keeping for post implementation defect analysis and debugging.
- Automated risk reports of daily cash letters for MUB Controllers on a quarterly basis.
RESEARCH
SOCIAL MEDIA AND INVESTOR RETURNS: THE CASE OF REDDIT- CSUF
August, 2016 - April, 2018
Coauthor: Prof. Arsenio Staer.
- We document a positive association between signed stock-related social media activity on Reddit and next day stock returns using a battery of sentiment analysis tests with a variety of controls.
- We used Python for data scraping and sentiment analysis.
- Using Python we scraped and filtered over 200 million Reddit posts based on stock tickers mentioned.
- Using SQL and SAS Enterprise Guide we created comparison tables with CRSP data and ran regressions.
- We ran sentiment analysis using the Loughran-McDonald and a modified VADER methodology.
- SSRN Abstract: https://papers.ssrn.com/sol3/papers.cfm?abstract_id=3282828
EDUCATION
BACHELORS IN BUSINESS ADMINISTRATION- CSUF
August, 2016 - April, 2019
- Minor in Finance.
- Financial Officer of the TitanRover engineering club.
- Minored in Finance with a focus on financial markets.
- Won 2nd place in a 32nd annual California State University Student Research Competition for presenting financial research.
- Raised $20,000 in donations as the Financial Officer for the TitanRover engineering club
