Home / Writing / Books / Resume Download PDF
Phillip Chaffee, Founding Software Engineer
| 443-553-5272 | San Francisco, US | LinkedIn | Github
SUMMARY
Software engineer with 10+ years of experience building distributed backend systems. Currently founding engineer at Sela AI, building conversational AI phone agents across the voice runtime, backend, telephony orchestration, Kubernetes scaling, and observability. Strong system-design bias; happiest collaborating across product, infra, and ML.
SKILLS
Backend: Python, Django, FastAPI, Node.js/Typescript, C#, Ruby, Celery, RabbitMQ, Pipecat, Twilio, gRPC
Data: PostgreSQL, Redis, SQL, NoSQL
Frontend: React, Angular
Infrastructure: AWS, Docker, Kubernetes, Helm, Git, Linux, Terraform, Statsig
Observability: Grafana, Prometheus, Loki, Tempo, Sentry
AI Development Tools: Cursor, Claude Code, GitHub Copilot, MCP, Cursor Rules
EXPERIENCE
Founding Software Engineer - Sela 2024 - Present
Python, Django, FastAPI, PostgreSQL, Redis, RabbitMQ, Celery, AWS, Docker, Kubernetes, Helm, Twilio, Pipecat, gRPC, Terraform, Statsig
  • First engineer and first employee at an AI startup building conversational AI phone agents; shipped 430+ merged MRs across the founding Python microservices stack (Django, FastAPI, Celery, gRPC) on Kubernetes in 16 months, owning the founding stack end-to-end.
  • Designed the warm-transfer phone calling system that hands live calls between AI agents and human operators: built a shared orchestration library with typed parent/child state machines, compare-and-set transitions, and snapshot/diff merge for concurrent saves on a Redis-backed session store, then rolled it out across the voice runtime, backend, and a parallel e2e suite behind feature flags.
  • Owned Kubernetes scaling and cost posture for the voice runtime: built predictive autoscaling on KEDA driven by scheduled-call counts 15 minutes ahead, replaced a 30-minute preStop sleep with a SIGTERM-aware graceful WebSocket drain across 4 gunicorn workers per pod, and cut overnight floor pods by ~75% with no impact on morning ramp.
  • Built the observability platform from scratch using Grafana, Prometheus, Loki, Tempo, and Sentry; instrumented call-quality and pipeline-timing histograms across services, then cut histogram cardinality by an order of magnitude by pruning high-cardinality labels when TSDB queries were timing out the on-call dashboard.
  • Shipped customer-facing telephony and AI-agent product features: call-outcome modeling driven by carrier disconnect reasons and Twilio AMD, agent / user talk-time analytics, retry and fallback dialing policies, AI agent tool-chaining, calendar-availability tools, and a system for rapidly cloning and configuring agents to accelerate customer onboarding.
  • Drove AI share of committed code to 87% across the engineering org (Cursor analytics, all repos) by authoring shared Cursor rules, a context-monorepo of agent skills (pre-MR checklist, design-doc rule, logging conventions), and rolling out Renovate, Ruff, and typing standards across every Python repo.
Senior Software Engineer - VoiceOps 2021 - 2023
Rails, React, Node.js, Python, AWS, PostgreSQL, Redis
Led engineering ceremonies and collaborated with senior engineers on new architecture including API design, microservice development, and an overhaul of the transcription platform.
  • Redesigned the transcription data pipeline, reducing cost by 90% and bringing cost per audio minute from $.20 to $.02.
  • Architected a call audio integration API and a microservice for text-audio alignment.
  • Implemented SAML SSO allowing customers to authenticate using their own identity provider. Used by 20% of clients.
  • Fulfilled on-call duties, resolving 5+ production issues each month.
  • Wrote a Typescript SDK for the AssemblyAI API.
Software Engineer - VoiceOps 2020 - 2021
Rails, React, Python, AWS, PostgreSQL, Redis
Built data collection platforms, processing pipelines, and customer-facing tools for a voice analytics startup.
  • Built a mechanical turk style transcription marketplace supporting 10k+ daily users performing manual transcription, transcript quality grading, and annotation tagging tasks.
  • Developed a data processing pipeline that transformed raw audio into transcribed and tagged text across multiple processing stages.
  • Built the customer-facing review platform for call center managers to view transcripts, review annotation labels, and provide coaching feedback.
  • Developed multiple customer integrations that imported data from external partners into the processing pipeline.
Software Engineer - Chatham Financial 2017 - 2020
C#, .NET, .NET Core, Angular, React, Python, SQLServer, Redis
Platform Engineering team building shared infrastructure used by development teams across the company: REST API templates, document storage, authentication services, and permissions APIs.
  • Built a .NET Core authentication microservice replacing a legacy library, supporting 10k+ requests/sec across all web applications.
  • Implemented SAML SSO for the flagship website, used internally and by multiple clients.
  • Migrated document storage from a distributed file system to blob storage with failover replication, serving all internal and client-uploaded documents.
Software Test Engineer - Chatham Financial 2016 - 2017
Ruby, C#, SpecFlow, RSpec, Selenium, Angular, MongoDB
Helped guide a small team of developers in creating the correct software for complex hedge accounting situations. Learned the hedge accounting system and constantly asked pointed questions to help guide the team’s thinking.
  • Collaboratively lead a project to create an internal test dashboard that aggregated every software team's test results into a single platform. The platform allowed users to view their test results including screenshots of failed UI tests and re-run the tests, no matter the test runner, directly from the platform.
  • Wrote automated acceptance/regression tests in C# and Ruby to confirm the software’s functionality daily.
PROJECTS
Built a virtual 16-bit Von-Neumann computer from N-And logic gates.
AWARDS
Gold Medal of Achievement (Eagle Scout), Royal Rangers 2013
Earned the Gold Medal of Achievement in Royal Rangers, a program that mirrors boy scouts. The GMA has all the same requirements as the Eagle Scout Award.