Fullstack Developer

Joined MeliorAI to contribute to the backend of its contract intelligence application, designed to analyze and extract key data and insights from legal documents while providing a ChatGPT-like question-and-answer feature amongst many other client driven needs.

Contributed to the development of a distributed workflow orchestration system designed to decouple client-specific business logic from APIs, ensuring robustness and flexibility across multiple use cases. Built using Ray for resource orchestration, Prefect for workflow orchestration, and Kubernetes for container orchestration. This three-layered approach ensures scalability, fault tolerance, and optimized resource utilization. The resulting data flows decouple client-specific business logic from APIs, enhancing flexibility for diverse use cases.

The company enlisted his assistance to migrate a monolithic SaaS app to the cloud. His responsibilities included developing a scalable infrastructure and transitioning app components into microservices;

Key Projects:

Rules Engine: Developed the company's pricing decision engine frontend using Django and JS. By separating pricing rules from the user codebase and moving them to a version-controlled ".json" config file in their AWS repository, dynamic rule building was enabled without client-side code changes. This facilitated continuous integration and deployment of pricing rules with zero downtime;

Payment Platform: Implemented the payment service, both frontend and backend, using the Stripe API. Overcame challenges like 3D secure checks, payment retries, trial periods, and coupons for smooth payment processing. Extended the app to integrate with Shopify and Magento, with a flexible and compatible frontend for both APIs;

Self-Service Reports: Developed a service enabling users to generate custom analytic reports. Integrated with the Google Drive API for report export to Google Drive and utilized Pandas for extensive dataset management and Elasticsearch and Kibana for data visualization and verification;

Google Integration: Created a web app enabling users to push specific data fields from their database to the Google Shopping and Shopify APIs. This functionality allowed users to annotate products on Google Shopping or update prices in their Shopify stores;

Testing & Code Style: Led team training on testing practices, including unit testing, PyTest, and mocking. Emphasized adherence to coding style guidelines like PEP8, linting, and standardized coding styles such as Black for improved code quality and reduced distractions;

Other: Worked on miscellaneous tasks such as backend fixes, legacy refactoring, frontend enhancements using CSS and HTML, and improving frontend aesthetics with the Keen Theme;

As part of the Production Engineering team, his primary focus was on designing, developing, and testing microservices and platforms to support the compute workloads of various engineering groups. They prioritized reliability, availability, and resource optimization.

Key Projects:

Containerizing The Grid: Took the lead in deploying Docker with Harbor onto the company's in-house grid network. This implementation enhanced the portability and flexibility of internal user flows, enabling platform-agnostic usage and laying the foundation for the future extension onto the cloud;

Diskovery: Contributed to the development of an internal solution for disk space management. The tool utilized multi-processed disk crawling techniques and provided lightning-fast querying capabilities. It outperformed available tools like du, duc, and ncdu, enabling efficient analysis of large-scale storage systems;

Automatic Code Reviewer: Developed a Python-based "bot" to assist with company code reviews. Leveraging Jenkins, the bot scheduled and executed various stages of validation, including static code analysis, compliance checks, integration testing, and unit testing;

In addition to these projects, he actively participated in conducting blameless post-mortems following production incidents, adhering to SRE principles. Additionally, communicated key metrics and incident insights to stakeholders;

As a Junior Tech Lead in the Verification team of the PowerVR Vision & AI department, his primary focus was on integration and feature testing of the company's NNA product. He was also a key contributor to the team's in-house automated CI/CD infrastructure.

Key Projects:

Plugin System: Led the planning and development of a testing framework that provided a CI/CD service for every engineering team in the department. The framework, written in Python, featured a product-agnostic Object-Oriented API with a modular design. Its flexibility allowed for easy addition or modification of components to support various test harnesses. The AI Engineering group successfully adopted this framework, utilizing it for numerous releases since 2018;

NNA/ISP Verification & Validation: Developed top-level integration tests for both the software stack and hardware simulation of the company's NNA and ISP products. Utilized industry-standard machine learning libraries such as TensorFlow, Caffe, ONNX, and TVM to generate reference results for performance and quality comparisons against the company's own solutions;

In these projects, he played a crucial role in ensuring the functionality, performance, and reliability of the NNA product through thorough integration testing and collaboration with machine learning libraries;

• Developed an automated video stream importer in Python (Django) that used the company's proprietary decoder to extract the information and upload it into a MySQL database;