Saad Khan

Frequency:

• Designed and implemented Pixel Server, an event-tracking system for ad campaigns using VAST URLs.
• Developed no-code architecture utilizing AWS services for efficient data processing and scalability.
• Captured every request log via Application Load Balancer (ALB), storing logs in Amazon S3 for further processing.
• Automated AWS Glue jobs to aggregate event data (clicks, impressions, and start events), storing results in DynamoDB for fast querying and reporting.
• Used EMR Cluster, Spark, and PySpark to reprocess parquet files generated from request logs.
• Upgraded Nginx and Fluentd versions in the existing architecture for better log management.
• Integrated Step Functions and SQS Lambda to convert logs into CSV files and then into Parquet files, making them queryable via Athena.

High Rewards:

• Architected a Retrieval-Augmented Generation (RAG) system on AWS to automate data extraction, verification, and querying of retail sales receipts using LLMs.
• Integrated OmniParser wrapped in FastAPI to extract structured data from scanned or photographed receipts.
• Developed a custom-trained model to verify receipt legitimacy and detect fraudulent or malformed documents using key heuristics and patterns in text and metadata.
• Employed OpenCV to normalize images (rotation, denoising, and alignment), with early rejection of visually corrupted or incomplete receipts to maintain data quality.
• Enabled natural language querying on extracted receipt data using LangChain integrated with a retrieval backend and a fine-tuned language model, allowing users to query using human language (e.g., “Show me purchases above $50 in March”).
• Built a rules engine to assign reward points based on receipt content (product type, amount, vendor), supporting real-time loyalty logic.
• Hosted the pipeline on AWS, utilizing Lambda, API Gateway, S3, and Step Functions for a serverless, scalable deployment.
• Created detailed technical documentation and diagrams to support onboarding, audit trails, and ongoing maintenance.

Bokksu:

• Led the architecture optimization of Bokksu's data pipelines across multiple Shopify stores integrated with Amazon, Walmart, TikTok, and other sources.
• Built ETL pipelines using webhooks, API Gateways, and AWS Lambda to collect raw data and store it in S3.
• Designed a data pipeline that moves data from S3 to Snowflake using Snowpipe, where DBT transforms and loads data into final tables.
• Integrated Mode Report and Zenlytics, enabling the marketing analytics team to generate reports using prompts.
• Implemented Matomo for user fingerprinting.
• Created GCP pipelines using Document AI, buckets and AppScript to extract shipping invoice details.

• Company Overview: Redwood City, CA
• Developed and optimized automation frameworks for large-scale applications.
• Integrated cloud-based services and machine learning models for enhanced testing automation.
• Redwood City, CA

• Company Overview: Palo Alto, CA
• Developed a data-driven automation framework for testing web and mobile applications.
• Created automated testing pipelines integrating APIs such as TestRail, DataDog, and Slack.
• Designed and built Ciitizen-e2e-UI, a web-based testing orchestration tool using Angular and Python Flask.
• Implemented Kubernetes-based job management for testing automation, streamlining deployment and execution.
• Palo Alto, CA

• Integrated API endpoints from various third-party services for improved functionality and feature-rich applications in .NET environments.
• Mentored junior developers, sharing knowledge of best practices in .NET development and full-stack technologies.
• Coordinated with product managers to gather requirements, create technical specifications, and develop new features accordingly.
• Improved application security by identifying vulnerabilities during code reviews, implementing robust encryption techniques, and staying current on industry best practices regarding cybersecurity measures in .NET development projects.

• Company Overview: Lahore, Punjab
• Developed Java Automation Framework (JAF), a data-driven automation solution for web and mobile testing.
• Built JAF Dashboard & Agent, a UI-based test execution tool integrating Jenkins, XenServer, and VMWare.
• Led test automation initiatives, reducing manual testing efforts and enhancing test coverage.
• Lahore, Punjab