Senior AI/ML Engineer & Backend Architect

Patrick Sun

AI / ML Engineer Senior Software Engineer

*************@*******.*** +1-339-***-**** Dublin, CA 94568 LinkedIn SUMMARY

Senior Software Engineer with 9 years of experience building intelligent search systems, LLM-powered applications, and large-scale backend infrastructure. Experienced in designing Retrieval-Augmented Generation

(RAG) pipelines, semantic search systems, and AI agent tool APIs. Strong background in distributed systems, cloud-native architectures, and data pipelines using Python, Kubernetes, and AWS. Proven experience building production AI services handling large-scale enterprise data and powering intelligent knowledge discovery systems.

TECHNICAL SKILLS

Programming Languages & Frameworks

Python, Java, TypeScript, .NET, Go, Rust, C#, C++, SQL, Bash/Shell, FastAPI, Flask, Django, React.js, REST APIs, gRPC, GraphQL, WebSocket, Kafka, OAuth, OpenAPI/Swagger, Webhooks, Protocol Buffers Machine Learning & AI

LLMs, Retrieval-Augmented Generation (RAG), GenAI, LangChain, LlamaIndex, Prompt Engineering, Prompt Optimization, Function Calling, Tool Calling, Agent Orchestration, Multi-Agent Systems, Embeddings, Semantic Search, Dense Retrieval, Reranking, Vector Search (FAISS, Pinecone), ONNX Runtime, AI Agents Data Engineering & Analytics

Apache Spark, PySpark, Apache Airflow, Hive, Presto/Trino, Hadoop, ETL Pipelines, Data Processing Pipelines, Data Modeling, Feature Engineering, Statistical Analysis, pandas, NumPy, SciPy Databases & Search Platforms

PostgreSQL, Redis, Elasticsearch, Cassandra, MongoDB, SQLAlchemy, Vector Databases (PGVector, FAISS, Pinecone), OpenSearch

Cloud & Infrastructure

AWS (EC2, S3, Lambda, SQS, RDS), GCP, Azure, Docker, Kubernetes, Terraform, Linux Systems, CI/CD (GitHub Actions, Jenkins, GitLab, CircleCI), Distributed Systems, Event-Driven Architecture, Async Processing MLOps, Observability & Experimentation

Model Deployment & Serving, Model Versioning, Model Registry, Online/Offline Evaluation, A/B Testing, Drift Detection, MLflow, TensorBoard, Observability (Datadog, Grafana, OpenTelemetry), Tracing, Logging, Metrics Dev Tools & Testing

Git, GitHub, pytest, Regression Testing, Golden Datasets, Offline Evaluation, Offline Replay Testing, Snapshot Testing

PROFESSIONAL EXPERIENCE

Senior Software Engineer — AI/ML Dropbox Oct 2021 – Present Dropbox Dash — AI-Powered Universal Search & Knowledge Discovery Backend engineer on Dropbox Dash (launched June 2023), an enterprise AI search product that connects company knowledge across files, messages, and apps in one place. Worked on the retrieval, indexing, and AI answer systems that make search smarter and responses more accurate.

• Built retrieval services on top of Dropbox’s internal search platform (Nautilus), contributing to query parsing, ranking, and indexing workflows using Python and gRPC to surface the most relevant content across connected apps like Slack, Google Workspace, and Dropbox.

• Worked on Dash’s RAG (retrieval-augmented generation) pipeline — the system that retrieves relevant company documents and feeds them as context to an LLM to generate accurate, grounded answers to user questions.

• Contributed to multimodal indexing pipelines that process text documents, images, and video transcripts into shared vector representations, enabling search across different content types and file formats.

• Developed query parsing and result ranking services that interpret search intent and apply multi-phase retrieval logic to surface the most relevant results with low latency.

• Built backend tool APIs used by Dash’s AI agents to fetch relevant company content on demand during reasoning, enabling more accurate and grounded answers.

• Contributed to Dash’s MCP (Model Context Protocol) tool layer, enabling AI agents to securely search and retrieve company content from within external MCP-compatible apps.

• Improved retrieval performance by implementing Redis caching and optimizing PostgreSQL metadata filtering strategies to reduce latency across large enterprise content collections.

• Supported production reliability of retrieval services using Datadog, Grafana, and OpenTelemetry for monitoring, alerting, and distributed tracing.

• Deployed and maintained services using Docker and Kubernetes, with canary deployments and automated rollback to protect search reliability.

Dropbox Paper — Collaborative Document Platform

Worked on backend document search and discovery services within Dropbox Paper workspaces, before moving to the Dash AI team.

• Developed Python APIs using GraphQL and REST for query parsing, document filtering, and result ranking within Paper workspaces.

• Improved search speed by optimizing PostgreSQL queries and indexing strategies for large team workspaces.

• Reduced database load by implementing Redis caching for frequently executed search queries.

• Expanded pytest test coverage for search logic using golden datasets and offline replay tests to catch regressions early.

Software Engineer — Cloud & Backend Infrastructure Amazon Web Services Jul 2016 – Jul 2021 Amazon Flow — Internal Workflow Automation Platform Owned the workflow execution service for an internal platform coordinating long-running backend operations and data processing jobs at Amazon.

• Built Python microservices and REST APIs for workflow state management, failure handling, and service orchestration.

• Designed a PostgreSQL persistence layer on AWS RDS for execution state and job metadata, ensuring work is not lost when services go down.

• Made failure-prone backend paths more resilient and resumable, reducing the need for manual intervention during job failures.

• Built deployment pipelines using Docker, Kubernetes, Jenkins, and GitHub Actions. AWS Elastic Beanstalk — Distributed Web Platform & DevOps Worked on background job infrastructure to separate long-running processing from user-facing application flows, improving reliability and response times for Elastic Beanstalk-hosted services.

• Implemented AWS Lambda workers and a Redis-backed task queue for scheduling and retrying background jobs.

• Separated async processing from user-facing flows, improving operational reliability and application responsiveness.

• Built containerized CI/CD pipelines with Docker and Kubernetes to standardize deployments and reduce production incidents.

Software Engineering Intern LG Silicon Valley Lab Jun 2015 – Aug 2015 Developed the Favorite Content application for LG’s 2016 Smart TV platform. Implemented front-end focus navigation logic and improved user interaction flows within the TV interface. EDUCATION

B.S. in Computer Science — University of California, Berkeley — 2013–2016

Contact this candidate