TianJian Wang

I’m a Senior Software Engineer focused on building scalable full-stack, cloud, and AI platforms. I’ve delivered enterprise-grade systems across Meta, AWS, and YourMechanic, with a strong emphasis on distributed systems, platform reliability, and developer experience.

At Meta, I led end-to-end development of Meta’s Llama developer platform—architecting full-stack AI infrastructure with React, TypeScript, Python, FastAPI, Kubernetes, Terraform, and GPU inference stacks. I built OpenAI-compatible SDKs, developer portals, CLI tooling, and API gateway services that reduced onboarding from days to minutes and grew beta developer adoption by 80%+.

I also engineered high-performance distributed inference and RAG/agent orchestration systems. Using Kubernetes, Triton, vLLM, PyTorch (including FP8 quantization), and multi-tenant routing, I achieved 80–150ms P95 inference latency and supported 2K–5K QPS, while RAG retrieval landed with sub-100ms overhead using Milvus, Weaviate, LangChain, and vector embeddings. On top of that, I implemented CI/CD and observability (canary releases, automated rollback, Prometheus/Grafana/ELK/Datadog) and improved incident resolution time by 40%.

Earlier at AWS, I built high-availability infrastructure for Managed Service for Apache Flink, delivering 99.99% uptime and reducing deployment-related failover latency by 40%. I also improved streaming performance (p99 < 500ms), enabled safer rollouts, and tightened monitoring/incident response with CloudWatch, while contributing internal and public documentation adopted by multiple AWS teams.