Kevin James

Staff Software Engineer | Lead Architect | Remote (CA/PT)

https://thekev.in
• (650) 686-9458
github.com/TheKevJames • linkedin.com/in/TheKevJames

Staff Engineer and Architect with 14+ years designing realtime, distributed systems for AI. Led multiple teams and acted as system architect responsible for all Voice AI systems handling 10B+ minutes per day peaking at 100M+ concurrent calls. Flagship work in building ASR and NLP systems and infrastructure on-par with Google for quality and performance at ~1/100th of the inference cost. Four patents in AI architecture, infrastructure autoscaling, and self-healing.

Skills

Infrastructure: GCP AWS Kubernetes (GKE) Terraform Docker Pub/Sub
Domains: Realtime Voice AI ASR/NLP Platform Engineering Event-Driven Architecture Observability Autoscaling & Self-Healing
Practices: System Design & Architecture Technical Leadership Cross-Team Roadmap Ownership CI/CD Incident Response

Work Experience

Staff Software Architect, Tech Lead (AI) Dialpad
San Francisco, CA; Remote
May 2018 • Oct 2026
dialpad.com
  • Redesigned the realtime voice pipeline for self-healing and autoscaling, raising availability from 98% to 99.995% and cutting outage recovery from ~6 hours to 12 minutes across 100M+ concurrent calls.
  • Architected and led the Voice AI platform serving 10B+ minutes/day at peaks of 100M+ concurrent calls, owning system design across ASR, NLP, and realtime telephony infrastructure.
  • Delivered ASR and NLP systems matching Google's quality and performance at ~1/100th of the inference cost, via custom infrastructure and compute optimization.
  • Grew the Voice AI organization from a 2-engineer team into 4 sub-teams of 6+ engineers each, personally mentoring the majority of the org - driving 9 internal promotions and 95%+ retention - and established a culture of retrospectives, fireside deep-dives, and cross-team knowledge sharing.
  • Led dozens of high-impact architectural migrations across hundreds of engineers and scientists - including AWS→GCP, Python 2→3, pip→Poetry, task queues/Redis Streams→Pub/Sub, and Pusher→Ably - reducing combined infrastructure spend by >90% while maintaining 99.99% availability throughout.
  • Owned technical roadmap integration across 8 product and platform teams, achieving 95% on-time delivery on major initiatives while driving cross-department stakeholder alignment and company-wide system design standards.
Senior Server Engineer, Tech Lead TalkIQ (acquired)
San Francisco, CA
Aug 2017 • May 2018
talkiq.com
  • Open-sourced projects with a combined >800 million downloads
  • Managed engineering due-diligence and systems integration leading to successful company exit (acquired by Dialpad)
  • Advocated for and implemented self-healing and autoscaling infrastructure with CI/CD, automated linting/testing/profiling, and Terraform IaC
Infrastructure Engineer TalkIQ
San Francisco, CA
Aug 2016 • Dec 2016
talkiq.com
  • Overhauled all pipelined services for migration from AWS to Google Cloud
  • Rolled out observability - logging, metrics, traces, and profiles - across all company systems
System Developer & Operations Engineer (Intern) AeroFS
Palo Alto, CA
Jan 2016 • Apr 2016
aerofs.com
  • Built and released multiple SDKs and apps for filesharing and unidirectional transfer
  • Implemented full natural language file search across the entire product suite
  • Optimized system build and packaging procedure for continuous delivery and failure recovery across OSX, Windows, iOS, and Android, including system-specific feature delivery
Data Scientist & Software Engineer (Intern) Accuvit
San Francisco, CA
Apr 2015 • Aug 2015
accuvit.io
  • Improved Word Error Rate of speech transcription systems by >10% absolute
  • Researched and developed automated core customer insight reporting for question detection and objection tracking
Server Developer (Intern) Enflick
Waterloo, ON
Aug 2014 • Dec 2014
textnow.com
  • Implemented automated ML systems for spam detection at scale
  • Optimized message recognition systems to handle millions of concurrent users
Lead Data Engineer (Intern) Wriber
Waterloo, ON
Jan 2014 • May 2014
wriber.com
  • Created NLP insight generation systems for helping writers expand and improve their work
  • Designed and developed self-learning ML systems for ranking, regression analysis, and unbound classification
Software Developer (Intern) Oracle
Cambridge, MA
Mar 2013 • Aug 2013
oracle.com
  • Wrote streaming JSON parser for converting highly nested input into standard RDB format
  • Automated notoriously difficult customer installation processes

Patents

Automated Pipeline for ASR speech modeling
Aug 2022
US Patent #11,410,658 B1
US Patent Office
Universally Proactive Compute Resource Scaling
Jan 2022
US Patent #11,226,844 B1
US Patent Office
Automated AI Model Development
Nov 2021
US Patent #11,169,798 B1
US Patent Office
Integrated AI and VoIP Platforms
Jun 2021
US Patent #11,050,807 B1
US Patent Office

Selected Writing

How Dialpad Moved Python AI Development from Pip to Poetry
Oct 2021
Syndicated to DevOpsChat, Python Weekly
TheNewStack
Kubernetes Engine (GKE) Multi-Cluster Lifecycle Management (multi-part series)
Apr 2020
Collaboration with GKE Architects to describe best-practice Kubernetes lifecycle management
Google Cloud Blog

Open Source Projects

gcloud-aio-* and gcloud-rest-*
Oct 2017 • Now
Python packages for interacting with Google APIs over public APIs
github.com/talkiq/gcloud-aio
  • >750M unique downloads, >10M downloads per day
coveralls-python
Aug 2017 • Now
Python coverage stats publisher for coveralls.io
pypi.org/project/coveralls
  • >75M unique downloads, used by >250k OSS projects
puppet-homebrew
Feb 2016 • Now
Homebrew provider for Puppet, used by most Puppet-managed macOS systems
forge.puppet.com/modules/thekevjames/homebrew
  • >350k unique downloads

Education

University of Waterloo
Sep 2012 • Apr 2017
Bachelor Software Engineering, Honours