软件工程师求职

Developed a support ticket triage system that processes text, image, and voice inputs using OpenAI Vision and Whisper APIs, enabling rich multimodal analysis.

Built asynchronous microservices in Spring Boot and Node.js to handle OCR extraction, speech-to-text conversion, and intent detection pipelines.

Indexed historical ticket embeddings using Pinecone and metadata filters to allow semantic similarity search and auto-routing.

Leveraged GPT-4 for sentiment scoring and urgency estimation, reducing manual triage load by 65% and increasing prioritization accuracy.

Designed Slack integration to auto-summarize tickets and suggest next steps, improving average first-response time by 42%.

Implemented a feedback loop for support agents, enhancing model relevance through RLHF-style prompt tuning.

Deployed system via AWS ECS with auto-scaling, and added OpenTelemetry monitoring for latency and failure tracking.

Ensured compliance with internal security policies by anonymizing sensitive fields and applying rate limiting on APIs.

Engineered a high-performance graphics engine using C++, integrating a SIMD math library, and achieving a 50% acceleration in runtime speed for matrix transformations and vector calculations through parallel computations.

Optimized the rendering pipeline, executing matrix transformations for dynamic graphics rendering, and developed custom shaders using HLSL, implementing advanced lighting models such as physically based rendering (PBR) and shadow mapping.

Designed and implemented a vertex management system from the ground up, enhancing the graphical rendering pipeline with efficient low-level buffer upload mechanisms and optimized data formats, reducing rendering overhead.

Improved rendering performance by 20.3% in scenarios with high character or object counts using multithreaded programming techniques, balancing workload distribution across CPU cores.

Enhanced system scalability and performance through advanced resource management techniques, including batch rendering and geometry instancing, which reduced draw calls by 35%, improving rendering throughput.

Collaborated with a team of designers and developers to align rendering engine optimizations with artistic requirements, ensuring seamless integration of advanced visual effects without compromising performance.

Worked closely with cross-functional teams to debug and profile the engine using NVIDIA Nsight and RenderDoc, identifying performance bottlenecks and implementing improvements that increased frame rates across various test scenarios.

Designed a custom memory allocator to improve memory access patterns, reducing memory fragmentation by 25% and improving overall system efficiency in real-time graphics applications.

Built a dynamic feature flagging system that uses GPT-4 to evaluate rollout logic based on user roles, regions, and behavioral signals.

Created a rule-based templating engine allowing product teams to define prompt-driven conditions in JSON, supporting 25+ live configurations.

Stored flag state and logic in PostgreSQL with Redis caching, reducing evaluation latency from 80ms to 25ms under production load.

Integrated gRPC server for runtime toggle evaluation with safe-mode fallback, improving rollout stability for dependent services.

Designed a React dashboard with preview simulation and analytics, enabling non-engineers to validate logic before deployment.

Enabled staged rollouts and tracked exposure metrics to support A/B testing scenarios.

Collaborated with SRE and product teams to define incident fallback strategies and automated toggle reversion.

Collaborated with engineers, product managers, and UX designers to build an Internal Goods Sharing and Collaboration Platform, enabling employees to share and interact through comments, images, videos, and other content. Utilized React, Spring Boot, Redis, MySQL, and AWS to deliver a seamless user experience.

Developed responsive and interactive frontend interfaces using React, HTML5, CSS3, and Bootstrap, while implementing backend services with Java Spring Boot, supporting complex business logic and RESTful APIs.

Designed and implemented a global login validation interceptor, reducing redundant code by 30% and enabling dynamic request path registration for flexible and secure authorization control.

Deployed the application on AWS EC2 instances with automated server health checks, ensuring secure, scalable, and highly resilient hosting.

Leveraged Amazon RDS for cloud storage, optimizing CRUD operations and achieving 3x faster query performance by applying advanced indexing and query optimization techniques.

Implemented Elastic Load Balancing (ELB) to distribute incoming traffic across multiple EC2 instances, supporting up to 200,000 read QPS, and maintaining application performance during high traffic periods.

Enhanced caching strategies with Redis, reducing backend response times by 35%, and improving system throughput to handle higher loads effectively.

Strengthened security and scalability by integrating IAM roles and fine-grained access controls, ensuring robust data protection.