Experience
2024 — Now
Seattle, Washington, United States
Developed and deployed AI-driven recommendation models using deep learning and collaborative filtering to enhance Snapchat Discover and Spotlight content suggestions. Implemented real-time personalization algorithms, increasing user engagement by 35% through vector search and retrieval-augmented generation (RAG) techniques.
✅ Programming & Frameworks: Python, Scala, Java, FastAPI, Flask, Django, PyTorch, TensorFlow, Scikit-Learn
✅ AI/ML & Deep Learning: LLMs (GPT, BERT, CLIP), NLP, Computer Vision (YOLO, OpenCV, Detectron), Recommendation Systems, GANs, RLHF
✅ Big Data & Real-Time Processing: Apache Spark (PySpark), Kafka, Snowflake, Airflow, Delta Lake, Redis, Elasticsearch
✅ Cloud & MLOps: AWS (SageMaker, Lambda, Glue, Bedrock), GCP (Vertex AI, BigQuery), Kubernetes, Docker, MLflow, Terraform, CI/CD Pipelines
✅ Databases & Storage: PostgreSQL, MySQL, MongoDB, DynamoDB, FAISS, Pinecone, MinIO
Developed real-time NLP-powered ad targeting models using Transformer-based architectures (BERT, GPT), increasing ad personalization effectiveness. Optimized AI model performance with ONNX, TensorRT, quantization, and pruning, improving inference speeds on mobile devices.
Architected and automated MLOps pipelines using AWS SageMaker, MLflow, and Kubernetes, reducing deployment times by 50%. Deployed serverless AI inference systems using AWS Lambda and FastAPI, optimizing API response times for real-time recommendations. Designed self-service AI experimentation environments with Jupyter, Zeppelin, and BigQuery, accelerating ML development workflows.
Built and managed real-time AI data pipelines using Kafka, Snowflake, Airflow, and Delta Lake, handling terabytes of user engagement data. Developed vector search and retrieval-augmented generation (RAG) techniques using FAISS and Pinecone for content discovery and ad targeting. Implemented event-driven AI architectures with AWS Kinesis and Kafka Streams, ensuring real-time content personalization.
2023 — 2024
2023 — 2024
Developed and maintained scalable backend services using Python (Flask, FastAPI) to support Snapchat’s core features, ensuring high performance and low latency for millions of users globally.
Optimized data processing pipelines with Python (Pandas, NumPy) for handling large-scale user data, improving the efficiency of data retrieval, transformation, and loading (ETL). Designed and deployed cloud infrastructure on AWS (EC2, S3, RDS, Lambda) to support scalable, fault-tolerant applications, leveraging auto-scaling and load balancing to handle fluctuating traffic.
Utilized AWS Lambda for serverless computing, reducing infrastructure overhead and ensuring fast and reliable execution of Snapchat’s microservices and event-driven workflows.
Built and maintained data lakes using AWS S3 and Glue, optimizing storage and data processing for large volumes of user-generated content. Developed RESTful APIs using Python (Flask/Django) to facilitate seamless integration between frontend applications and backend services, ensuring efficient data exchange for Snapchat’s core functionalities.
Integrated third-party APIs to enhance platform features, including image and video processing, ad services, and social media sharing. Implemented CI/CD pipelines using AWS CodePipeline, CodeBuild, and Jenkins to automate testing, integration, and deployment of Python-based applications, reducing deployment times and improving code quality.
Deployed containerized applications using Docker and Kubernetes on AWS EKS, ensuring high availability, scalability, and fault tolerance in production environments. Built ETL pipelines using AWS Glue and Python for processing and transforming large datasets, enabling efficient ingestion of real-time data from multiple sources into data lakes and databases.
Developed real-time analytics systems to track user engagement, using Python and AWS services (Kinesis, Redshift) for large-scale data processing and real-time insights into user behavior.
2022 — 2023
2022 — 2023
Used AWS CloudFormation to provision, update, and manage a collection of related AWS resources in a consistent and controlled manner. Used CloudFormation template with DSOP to deploy the application into a test environment created explicitly for security validation and vulnerability scan. Used DSOP as a Jenkins pipeline within Amazon EC2, with Amazon S3 used for log capture and other storage tasks also, used AWS Lambda for custom configuration rules. This helped company meet all security configuration rules also, enables the company to catch potential security risks prior to production deployment. Created Python programs to increase efficiency of application system & developed Python Framework using Django to perform scan software unit monitoring. Developed User Interface using technologies like HTML, CSS. Developed server-based web traffic statistical analysis tool using Flask, Pandas.
Managed all Serverless functions with the Serverless Framework allowing for ease of management and cloud provider flexibility also developed and implemented secure AWS Lambda serverless functions in Python.
Implemented a 'server less' architecture using API Gateway, Lambda, and DynamoDB and deployed AWS Lambda code from Amazon S3 buckets. Created a Lambda Deployment function and configured it to receive events from your S3 bucket.
Wrote python scripts using Boto3 to automatically spin up the instances in AWS EC2 and OPS Works stacks and integrated with Auto scaling to automatically spin up the servers with configured AMIs. Also, Using Chef, deployed and configured Elasticsearch, Logstash and Kibana (ELK) for log analytics, full text search, application monitoring in integration with AWS Lambda and CloudWatch.
Wrote and executed several complex SQL queries in AWS glue for ETL operations in Spark data frame using Spark SQL. Also, successfully migrated the Django database from SQLite to Postgres SQL with complete data integrity.
2021 — 2022
2021 — 2022
Developed Python based API (RESTful Web Service) to track the events and perform analysis using Django. Collaborated Flask and pandas to monitor, migrate and develop table database and work with large data set files. Designed the front-end applications, user interactive (UI) web pages using web technologies like HTML, CSS and Django.
Worked with container-based deployments using Docker, working with Docker images, Docker Hub and Docker registries and Kubernetes. Developed GUI using webapp2 for dynamically displaying the test block documentation and other features of Python code using a web browser.
Developed Microservices by creating REST APIs and used them to access data from different suppliers and to gather network traffic data from servers. Used Jenkins pipelines to drive all microservices builds out to the Docker registry.
Consumed external APIs and wrote RESTful API using Django REST Framework and Angular. Eqully comfortable working within the Django ORM or writing native SQL in SQL Server.
Installation, administration, and maintenance of CI/CD applications- Jenkins, Puppet, and Docker for designing and developing microservices and continuous deployment. Worked in an agile development environment. Also, Managed the Docker container through the pods and performed the load balance between the pods through Kubernetes.
Used Kubernetes to deploy scale and load balance, and worked on Docker Engine, Docker HUB, and Docker Images. Utilized Kubernetes and docker for the runtime environment of the CI/CD system to build, and test deploy.
Formulating the ETL mappings to implement the business logic. Used transformations like lookup, update strategy, expression, filter, router, aggregate, source Qualifier.
Exposure about ETL batch and the concept of data warehousing. Worked with container-based deployments using Docker, working with Docker images, Docker Hub and Docker registries and Kubernetes.