July 2024 Team Talk
August 1, 2024
Comments Off on WebRTC.ventures Team Talk: July 2024

Each month, we ask our dynamic team of WebRTC developers, mobile developers, designers, testers, DevOps experts, and project leads to share some of the impactful work they did in the past month.

Here’s a short list of some that might interest you — at least the ones we can talk about!

WebRTC

  • Post-Call Transcriptions: Integrated post-call transcriptions into a WebRTC application using Azure Speech SDK, enhancing the call review process.
  • WebRTC App Troubleshooting: Troubleshot connectivity issues in a Windows SDK WebRTC app caused by DTLS incompatibilities, ensuring smoother and more reliable connections.
  • AWS Chime Archive Video Retrieval: Working on obtaining unmixed, individual stream archive videos from AWS Chime meetings, working towards better meeting documentation and review.
  • Memory Issue Resolution: Resolved an issue with an app crashing due to large data buffers (videos converted to strings) exceeding memory limits.
  • Call Quality Assessment: Assessed call quality using inboundRTPStats, roundTripTime, and outgoingBitrate to inform users about call performance.

Conectara

  • The month saw the official launch of Conectara, our solution/service for modernizing contact center operations, powered by Amazon Connect. This project is truly a team effort!
  • Outbound Contact Flows Development: Developed outbound contact flows for a call center service using Amazon Connect, Amazon Lex, DynamoDB, and AWS Lambda functions. This has significantly improved our client’s call center operations.
  • Conversational Bot Research: Researched tools in AWS to integrate Amazon Connect with a conversational bot, paving the way for more advanced customer service solutions.

AI and Machine Learning

  • AI Assistant Bot Development: Developed an AI assistant bot for call centers enabling natural and context-aware conversations with customers.
  • AI Bot Service Cost Estimation: Estimated costs for a real-time LLM AI bot service using Amazon Lex, Amazon Bedrock, and Amazon Polly, preparing for future implementations.
  • AI Code Assistant Setup: Set up a local AI code assistant in VS Code to boost our development efficiency.
  • Voicebots and Dynamic Interactions: Created voicebots using Amazon Lex and integrated them with Amazon Bedrock for more seamless and dynamic interactions.
  • Local LLMs Experimentation: Experimented with local LLMs (Llama 3 and 3.1, Gemma 2, Phi 3, and Mistral AI) using Ollama and Open WebUI, keeping us at the cutting edge of AI technology.
  • Call Compliance Enhancement: Enhanced call compliance for customer service representatives using Amazon Transcribe for transcription, Amazon Comprehend for sentiment analysis, and Amazon Bedrock for generating call summaries and reviews.

DevOps

  • Application Environment Setup: Established new environments for applications, ensuring our development and deployment processes remain smooth and efficient.
  • Kubernetes Cluster Node Addition: Added a Kubernetes cluster node using Kubespray after decommissioning an old virtual machine, ensuring our infrastructure remains robust and scalable.
  • Developed and deployed Lambda functions using Serverless Framework (SLS) and AWS CloudFormation
  • Dead Letter Queue Implementation: Implemented redrive policies in Amazon SQS to manage undeliverable messages, utilizing Dead Letter Queues (DLQs) for improved troubleshooting and efficiency.

New Content

We released the following blog posts:

And these videos:

At WebRTC.ventures, we continually enhance our capabilities to provide the best real-time video, audio, chat, AI, and contact center solutions. Whether you need an assessment of your product or assistance with building, integrating, testing, and managing it, we are here to be your partner.

If you or anyone in your team is facing challenges building real-time communication applications or integrating AI, our team loves to find and solve any kind of problem. Contact us and let us know how we can help!

Recent Blog Posts