Senior Software Engineer

@ Avoma

April 2022 - Present

Built a secure, time bound access tool for support teams to inspect customer accounts.
Enabled independent investigation without engineering involvement, cutting nearly all exploratory escalations.
Improved auditability through reason logging and token expiration.

Improved and expanded integrations across Salesforce, HubSpot, Pipedrive, Copper and Outreach.
Delivered a unified sync interface that normalized CRM specific behaviors so the AI note system could work consistently across platforms.
Designed and implemented multi entity note sync, letting customers push meeting notes across multiple related CRM objects, significantly improving their in CRM visibility of Avoma data.
Reduced friction from CRM nuances by building abstractions that simplified client experience while preserving data integrity.

Rebuilt the bot architecture using Nodejs with a modular client server design that allowed multiple meeting platforms to be added and maintained independently.
Introduced a local testing workflow through Docker that replaced the slow remote testing loop, cutting testing time from 30 minutes to near instant.
Reduced build times from 30 to 60 minutes down to approximately 7 minutes using multi stage builds.
Improved bot join latency from around 30 seconds to under 10 seconds.
Led the full migration from the legacy system. After the first client shipped in 6 months, the team integrated all other meeting platforms within a year and fully retired old infrastructure.
Took ownership of the recording bot domain and mentored more than four engineers on design, debugging and platform extensions.

Implemented Zoom’s RTMP based streaming capability to support live transcription and faster post meeting processing.
Built scalable ingestion using nginx rtmp and HAProxy, enabling reliable handling of over one thousand hours per month.
Reduced overall meeting processing times by roughly 30 percent by switching to near real time ingestion instead of waiting for post meeting artifacts.

Technical Expertise

Backend Development: Extensive experience with Python and Nodejs for building scalable services, along with targeted use of C++ for performance sensitive components.
Frontend Development: Proficient in React for building product interfaces and Plasmo for developing browser extensions.
Infrastructure and DevOps: Hands on work with Docker based environments, Nginx and HAProxy for routing and load handling, ffmpeg for media processing, and tooling such as Packer, Vagrant and custom AWS AMI workflows to support reliable deployments and platform automation.

Professional Growth

Learned advanced system design and architectural decision making through close collaboration with highly experienced engineers.
Improved skills in scalability, observability, devops and high availability patterns.
Consistently led projects end to end, including planning, design reviews, execution and mentoring contributors.

Built a secure support access platform that eliminated engineering escalations for account investigations.
Enhanced CRM integrations across Salesforce, HubSpot, Pipedrive, Copper and Outreach with a unified sync framework.
Delivered multi-entity note sync enabling customers to map Avoma insights to multiple CRM objects.
Re-architected the recording bot system with modular Node.js architecture, reducing test cycle time from 30 minutes to near zero and build times from 30-60 minutes to ~7 minutes.
Improved bot performance by reducing join latency from ~30 seconds to under 10 seconds.
Scaling to 100,000 hours of recordings per month with greater than 99.95% availability and reliability.
Led full migration from legacy bot framework while mentoring 4+ engineers.
Implemented Zoom RTMP streaming for real-time transcription, cutting processing time by ~30%.