Senior Developer Ai Agent

US • 650-***-****

Xiang Ang

Professional summary

Senior Developer with extensive expertise in leading high-impact projects and optimizing system performance using cutting-edge technologies like Golang, NodeJS, and Terraform. AI part having deep experience in LLM/Deepseek/LangChain LangGraph Flowise MCP A2A/Onnx/Bert/SLM/YOLO etc for RAG usage. Demonstrated success in enhancing user engagement through innovative UI improvements and seamless project migrations. Passionate about leveraging advanced skills in Big Data and microservice architecture to drive future-ready solutions and foster a culture of collaboration and knowledge sharing.

Employment history

Senior Dev Jan 2023 - Jan 2025

GoogleUSA

Core Backend AI Agent Orchestration LLM Integration Cloud-Native Systems

●Led development of scalable backend services using Java (Spring Boot), Python (FastAPI, LangChain LangGraph Flowise MCP A2A agents), and Golang, supporting high-throughput APIs with REST, gRPC, and event-driven architectures.

●Designed and deployed LLM-integrated pipelines using LangChain LangGraph Flowise MCP A2A, Langflow, and ONNX-based local models (e.g., Deepseek, TinyBERT), powering AI assistants for internal analytics and secure document Q&A.

●Developed a modular AI Agent framework with context caching, multi-modal routing, and vector store memory (Qdrant/Weaviate), enabling agents to perform document parsing, API synthesis, and human-in-the-loop escalation.

●Optimized inference latency and model orchestration in on-premise environments, with fallbacks from cloud-hosted APIs (OpenAI, Gemini) to local transformers.

●Integrated AI-assisted SQL generation from natural language using fine-tuned Text-to-SQL models for internal teams—eliminating manual query writing.

●Built cloud-native microservices on GKE with Terraform, including CI/CD pipelines and service discovery, with resilience patterns using Envoy and Istio.

●Contributed to over 100+ PRs across services written in Java, Python, and Golang, improving observability, testability, and overall platform reliability.

Senior Developer Dec 2021 - Dec 2022

WalmartUSA

LLM-powered Search Kubernetes Microservices High-Performance APIs

●Developed a sidecar microservice in Golang with WebPush + BoltDB, improving real-time messaging in distributed Kubernetes deployments.

●Migrated legacy REST interfaces to GraphQL APIs, significantly reducing response latency and payload size for internal dashboards.

●Fine-tuned Solr-based search and implemented LLM-based ranking models (with embedding + RAG retrieval), increasing relevance in enterprise document search.

●Built scalable backends in Python (Django, Flask) and Java (Spring Boot), containerized with Helm and deployed via Kubernetes (EKS).

●Improved performance on SingleStore and Snowflake by profiling query plans and integrating AI-based indexing strategies.

●Integrated Langflow workflows into developer portals to allow no-code prompt orchestration and agent chaining.

●Deployed tracing and observability stacks (OpenTelemetry + Datadog) across AI-backed services, enabling fast root cause analysis in distributed systems.

Senior Software EngineerJune 2020 - Dec 2021

Mastercard · USA · Jun 2020 – Dec 2021

Real-Time Payments AI Agent Systems Secure APIs Multimodal LLMs

At Mastercard, I led initiatives to build secure, scalable, and intelligent financial platforms — combining real-time payment processing, AI-driven fraud detection, and regulatory-compliant APIs.

Real-Time Payment Systems

●Engineered high-throughput transaction services using Java (Spring Boot) and Kafka, handling millions of TPS with consistent low-latency.

●Integrated MPGS (Payment Gateway Services) with full PCI-DSS compliance, optimized ISO 8583 parsing, and built failover strategies on Kubernetes (AKS/GKE) + CockroachDB.

AI-Powered Fraud Detection & Risk Intelligence

●Designed AI/ML-based fraud detection pipelines using Java, Flink, Spark, and Drools, detecting anomalies in global transaction patterns.

●Integrated Mastercard’s Decision Intelligence API for adaptive, real-time risk scoring with reduced false positives.

●Deployed LLM Agents using LangChain LangGraph Flowise MCP A2A to extract and summarize suspicious transaction traces and flag them to human reviewers.

●Combined image-blind LLMs with structured tabular data to detect altered payment screenshots, scanned documents, and synthetic fraud attempts.

API Security & Compliance

●Developed REST/gRPC APIs for fintech partners, ensuring Open Banking + PSD2 compliance and SCA (Strong Customer Authentication).

●Built tokenization services for PAN data, leveraging HSM-based encryption (FIPS 140-2), and applied OAuth2 + MFA for access control.

●Introduced GraphQL endpoints for unified financial data access, improving internal and partner query efficiency.

Performance, Observability & CI/CD

●Optimized system performance via JVM tuning, GC profiling (JProfiler, VisualVM), and async I/O designs.

●Enabled autoscaling, circuit-breaking, and fallback strategies using Spring Cloud, Resilience4j, and Istio.

●Instrumented services with OpenTelemetry, enabling full-stack tracing, anomaly alerting, and audit-friendly observability.

Streaming & ETL at Scale

●Built real-time reconciliation services with Apache Flink + Kafka Streams, improving daily settlement timelines.

●Designed AI-assisted ETL pipelines using Apache Beam + Google Dataflow, analyzing billions of transactions for compliance and insights.

●Used BigQuery to power AI agent dashboards and feed vectorized data into embedding stores for cross-modal analytics.

Strategic Projects – Multimodal & Cross-Agent AI

●OpenStack Neutron at Scale: Rebuilt Neutron for 1M+ nodes, integrating ML anomaly detection for self-healing networking.

●AI Agent Log Intelligence: Developed LangChain LangGraph Flowise MCP A2A agents to scan system logs and correlate behavioral anomalies across services.

●Cross-Modal Fraud Detection: Integrated vision-capable LLMs (e.g., BLIP, MiniGPT) with transaction metadata to detect fraudulent receipts, screenshots, and manipulated financial docs.

●Langflow + Vector Store Agents: Deployed Langflow-based AI orchestration to compose retrieval, summary, and image classification tasks—bridging tabular, text, and visual data in unified flows.

●Predictive DevOps (Spinnaker OSS): Created AI-based resource prediction agents that tuned deployment configs dynamically based on usage trends and risk scoring.

Lead Software Developer and architect NOV 2019 - Jun 2020

UBERTAL INC Insurance company ContractorCA, USA

UBERTAL Inc. (Insurance Company Contractor) Nov 2019 - June 2020 USA

●Optimized Spring Boot projects for insurance financial systems, improving processing time from 40 minutes to 40 seconds (60x performance boost).

●Developed API monitoring tools in Java and Golang, identifying bottlenecks and optimizing database transactions and API response times.

●Led the migration of legacy systems from IBM SoftLayer to GCP, refactoring services for cloud-native deployment.

●Implemented distributed caching using Redis and Memcached, significantly reducing database load.

Senior Software Engineer

UBERTAL Inc. (Google PSO Contractor) May 2019 - Nov 2019 USA

●Led AI-driven performance tuning, using machine learning models to analyze API logs and improve response times.

●Optimized API latency by identifying and resolving performance bottlenecks in Java-based services during migration from IBM SoftLayer to GCP.

●Developed a Universal Scheduler using Golang’s goroutines, improving parallel execution by 50x.

●Integrated AI models for predictive scaling, ensuring dynamic resource allocation based on traffic patterns.

Senior Software Engineer

UBERTAL Inc. (Futurewei Contractor) Dec 2016 - May 2019 USA

●Developed a cloud-native multi-tenant JVM solution, saving 10x infrastructure costs by transitioning legacy microservices to Kubernetes.

●Reengineered microservices in Java, Golang, and Python, optimizing resource utilization and runtime performance.

●Built a real-time event-driven pipeline using Kafka, GCP Pub/Sub, and Flink, improving message processing efficiency.

Software EngineerJun 2016 - Nov 2016

Plume DesignPalo Alto, CA, USA

●Developed a high-performance HAProxy-based load balancer for AWS, Azure, and PCF, replacing AWS IoT services while achieving 20k TPS throughput.

●Optimized MQTT brokers for low-latency IoT message streaming, enhancing real-time communication for smart home devices.

●Implemented client certificate verification and OCSP Server for MQTT authentication, improving security in IoT networks.

●Performed in-depth performance analysis on MQTT and network protocols, optimizing message throughput and scaling strategies.

Lead Developer & Solution ArchitectOct 2014 - Jun 2016

Shaklee CorpPleasanton, CA, USA

Project #1: Automation & DevOps Infrastructure

●Designed and automated CI/CD pipelines using Vagrant, Packer, Docker, and Chef, streamlining infrastructure provisioning and application deployment.

●Integrated AI-driven code analysis tools to detect inefficiencies early in the CI/CD process, reducing build times and optimizing resource allocation dynamically.

●Collaborated with IBM vendor teams on Android/iOS app development, performing code reviews and AI-based continuous testing for code quality assurance and security.

●Developed DevOps automation scripts in Python and Bash, optimizing configuration management and deployment workflows.

Project #2: Adobe AEM - Shaklee Website Development

●Developed and optimized the Shaklee website using Adobe AEM (CQ5, AEM 6.0, 6.1) and Sightly, improving content delivery speed and user experience.

●Implemented AI-powered analytics tools (Omniture, Google Analytics) for traffic analysis, user behavior insights, and SEO optimizations.

●Designed RESTful APIs in Java and Python for content management and personalized recommendations based on AI-driven customer segmentation.

●Optimized backend performance by refactoring Java-based AEM components, reducing response times by 40%.

Project #3: Global Front Office Platform

●Led the successful integration of WeChat Pay and Alipay, ensuring secure payment processing with AI-driven real-time transaction monitoring and fraud detection.

●Developed scalable microservices on AWS using Java, Golang, HAProxy, and the Play Framework, improving system resilience and dynamic load balancing.

●Optimized API performance using AI-based predictive scaling models, reducing latency and improving response times for high-traffic payment transactions.

●Designed and developed frontend components using AngularJS and PubNub, implementing real-time communication and event-driven UI updates.

●Researched and implemented cloud-native solutions using Fabric8, Kubernetes, and CI/CD pipelines, improving deployment automation and fault tolerance.

Project #4: Daimler Project on Heroku

●Developed real-time data query services on Heroku using Golang, Angular, and Evothings, enabling seamless Elasticsearch integration.

●Applied AI-driven query optimization techniques to improve search performance, reducing latency and enhancing user experience.

●Built intelligent indexing mechanisms in Elasticsearch, leveraging machine learning models for predictive indexing and automated ranking adjustments.

●Optimized API interactions using Golang concurrency models, ensuring high-performance and scalable backend services.

Key Technical Skills & Hands-On Expertise:

●Backend Development: Java (Spring Boot, Play Framework), Python (Flask, FastAPI), Golang

●AI & ML: NLP (Text-to-SQL, GPT-based optimizations), Predictive Scaling, AI-driven Observability

●DevOps & Cloud: Kubernetes (GKE, AWS, Heroku), Terraform, Docker, CI/CD Automation

●Frontend Development: AngularJS, PubNub, AEM (Adobe Experience Manager)

●Search & Data Processing: Elasticsearch, AI-powered query optimization, Apache Kafka

Technical LeadFeb 2012 - Oct 2014

BOCOBEIJING, CHINA

Develop data-worker for our real time distribution system based on Teracotta Big Memory on huge amount of data using AKKA cloudy computing framework (JAVA/ADOBE AEM/Webstock/scala/groovy/kafka) and ML Shared Memory (os api) to solve the real time big data issue in telecommunication area.

Performance tune to find the performance issue, like high cpu usage, memory leak, IO bottle neck and network bandwidth, using tools like profiler, jdk commands and system commands.

Review the code and manage the case using Jira/Mantis

Create flexible project/development plan to meet changing needs and requirements of clients.

Coordinate with other teams (e.g support team) to handle urgent issue from customer when permit.

Team LeaderJan 2011 - Feb 2012

Talend Software CompanyBEIJING, CHINA

Develop MDM product using JAVA/ADOBE AEM/WEBSTOCK/GWT/GXT/EJB/WEB SERVICE/JBOSS/HIBERNATE/SPRINGP/PROTOTYPE/Kafka/JQuery/XPATH/JAXB/ python3.7

Analysis and performance tuning using tools (profiler, jdk command and system commands) to find bottleneck like high cpu usage, memory leak, IO and network bandwidth.

Manage and Assign the cases in the case management (Jira/Mantis). Help team members to solve problems.

Software EngineerOct 2009 - Jan 2011

RealNetworksChina

Develop BI Queue Reader and data transformer based on Teradata/Oracle using JAVA/ADOBE AEM/Webstock/Latest JDK based EDW/BI system (Zabbix+ Teradata + Oracle+SonicMQ) for Real Networks.

Design, implement, maintain and enhance legacy/new server-side software.

Fix performance issue using tools (profiler, jdk command and system commands) to find high cpu usage, memory leak, io bottle neck and network bandwidth.

Summarize, detail and deploy the module according to the requirements, write code and test.

Senior RV/EMS Kafka SupportApr 2006 - Sep 2009

TIBCO CDCBEIJING, CHINA

Write code to analyze the customer’s real production issue. Provide consultancy service to TIBCO CDC for TIBCO Rendezvous and Enterprise Message Service. Performance tuning to tune some performance issue using tools (profiler, JDK command and system commands) to find high cpu usage, memory leak, IO bottle neck and network bandwidth.

Provide training and help for PSG and other products team like BW, BPM, TIBCO-SDK etc.

Track completion of planned deliverables and reviews with manager.

Review code to ensure the TIBCO code quality standards; mentor junior support.

Education

Bachelor2002 - 2006

BJTU

Software Engineering

Master2008 - 2011

BUAA

Network Security – Software Engineering

Skills

Java (Expert), GenAI (Expert), GPT4, Llama3, HuggingFace, Hadoop, MapReduce, Flink, BigData, Nats.IO, Groovy, AEM, Angular, Softmax, PKI, Databricks, Apache Pinot, Trino, Scala, Kotlin, Micro Service, AWS, Azure, PCF, BlockChain, Confluent Kafka, GCP, BigQuery, CosmosDB, Data Pipeline, Splunk, Kubernetes, DevOps, ML, AirFlow, Go, Spark, Python, Reactjs, HTML, CSS, Vue, Crawler.

Languages

Mandarin (Native), English (Highly proficient), Japanese (Proficient).

Contact this candidate