Alexa Griffith

Summary

Software engineer specializing in large-scale AI inference systems and cloud-native infrastructure with a focus on building safe, observable, and evolvable AI systems in hybrid on-prem and cloud environments. I design and operate production GenAI platforms on Kubernetes, with experience in KServe, Envoy AI Gateway, and multi-cluster reliability. Active open-source contributor and CNCF community leader with multiple KubeCon/public talks and committee roles. Host and producer of Alexa's Input (AI), a podcast featuring engineers, founders, and open-source leaders.

Professional Experience

Bloomberg New York, NY

Senior Software Engineer - Inference Platform Team January 2022 - Present

• Lead engineer on Bloomberg's internal GenAI inference platform built on Kubernetes, KServe, and Envoy AI Gateway, powering production ML and LLM workloads across multiple clusters and environments.
• Designed and implemented core transformation logic in Envoy AI Gateway across providers, and multimodal/streaming/tool-calling use cases, enabling a single consistent API.
• Defined and implemented SLOs/SLIs and dashboards for the platform, deployment system, and user services creating Grafana dashboards and alerts used for ongoing reliability reviews and incident response. Led our Inference Platform Stability strategy.
• Owned multiple DR-1 disaster-recovery test cycles for inference, including rebuilding clusters, documenting procedures, and coordinating cross-team response. Served as the representative for my team.
• Designed and rolled out GPU support and configurations for inference services, including scale-to-zero.
• Setup AWS EKS-based inference clusters (first clusters on cloud); configured to be compatible with on-prem, onboarded the team to Terraform-based pipelines, and designed OPA workflows and policies.
• Led multiple KServe upgrade and rollout efforts (0.9–0.15) across shared and dedicated clusters.
• Authored upstream KServe features and documentation for GenAI runtimes, model latency metrics, and canary rollouts; maintained KServe implementation and website contributions.
• Built and designed the deployment and debugging backend platform service and integrated it with our UI: new APIs for pod status, health, and information, significantly improving debuggability for internal users.
• Drove cross-team initiatives with SRE, product, and other platform teams regularly.

Bluecore New York, NY

Software Engineer 1.2 - Data Science & Analytics Infrastructure Teams November 2019 - January 2022

• Decreased analytics UI response time by 95% (~2 minutes to sub-seconds) by leading a multi-quarter project for the design and development of a new gRPC-based Go API; main on-call for new & legacy systems.
• Pioneered the design and cross-team discussions for a standardized authentication and authorization service implementation for Bluecore's GKE-based microservices.
• Contributed crucial features such as database lock handling and a data generator service for our high throughput, low latency streaming architecture written in Python/Java.
• Created a SQL parser tool with a UI to visualize similarities in complex queries using d3/Python/JavaScript and consolidate 50% of the analytics endpoints.
• Managed Airflow infrastructure, added new features, and created optimized DAGs for ETL workflows.
• Led and designed Looker integration usage for Bluecore's UI used by 600+ clients & created self service features for users.

Open Source, Leadership & Activities

See full list at alexagriffith.com

Open Source

• Core contributor to KServe (GenAI/LLM support, metrics, Helm, docs) and Envoy AI Gateway (multi-provider schema mapping, streaming, tools, reference architecture).
• Active participant in Kubernetes, CNCF, and AI-platform working groups.

Speaking & Writing

• Conference speaker at: KubeCon EU/NA/JP/China, SRECon Dublin, KubeCrash, Grace Hopper and Bloomberg Developer Conference (BDC), with talks on GenAI platforms, KServe, SLOs/SRE for AI, and Kubernetes-based inference.
• Host of "Alexa's Input (AI)" podcast, interviewing engineers and open-source leaders (+50 episodes total; ~1 per month). Published on podcast apps, YouTube, and Substack blog. Notable guests include Kelsey Hightower, Joe Beda, Dan Walsh.
• Guest on MLOps and The Kubernetes Show podcasts; featured on MLOps newsletter.
• Authored:
  • Building a Scalable, Flexible, Cloud-Native GenAI Platform with Open Source Solutions
  • A Reference Architecture for Adopters of Envoy AI Gateway
  • Announcing KServe v0.15: Advancing Generative AI Model Serving

CNCF & Industry Roles

• Technical Advisory Group (TAG) Infrastructure – Tech Lead (CNCF).
• Program Committee Member: Cloud Native + Kubernetes AI Day EU 2026 & NA 2025, EnvoyCon (Virtual).
• Lead SME for Cloud Native Platform Associate Advisory Council; Advisory Panel Member for Cloud Native Platform Engineer Advisory Council.

Internal Culture & Mentorship

• Mentored interns and engineers interested in platform/inference work; ran debugging workshops and KServe training sessions across offices.
• Represent Bloomberg at recruitment events, Kubernetes meetups, and open-source showcases.

Technical Skills

• Languages: Go, Python
• Cloud & Infra: Kubernetes, KServe, Envoy AI Gateway, Istio, Prometheus, Grafana, Helm, Terraform, Docker, Knative
• Cloud Providers: AWS (EKS, Bedrock), GCP (GKE, Pub/Sub, BigQuery, Bigtable, CloudSQL, load balancers)
• Data & Streaming: BigQuery/Bigtable, PostgreSQL, Kafka

Education

University of Tennessee, Knoxville 2014 - 2019

B.S. in Honors Chemistry, American Chemical Society certified
Minor in Hispanic Studies
Magna Cum Laude