Author
Sonam Gupta, PhD
Developer Evangelist
Sonam is a San Francisco-based developer advocate, originally from India. She has completed 2 Master's Degrees and her PhD in Data Science from the Harrisburg University of Science & Technology. Previously, Sonam worked for the startups Ozmosi and aiXplain. In her free time, you will find Sonam dancing, writing, and exploring local coffee shops.
- 16
- Articles
- 2025
- Joined
Latest articles
Everything Sonam has published, most recent first.
InferenceGLM-5.3-Flash Benchmarks
GLM-5.3-Flash inference benchmarks comparing Telnyx, Baseten, Together AI, and Fireworks on throughput and completion reliability.
InferenceDeepSeek V4 Flash inference benchmarks on Telnyx
DeepSeek V4 Flash inference benchmarks comparing Telnyx StatefulActor, Baseten, Fireworks, and Together AI on E2E latency and throughput.
VoiceWhat Developers Learn After Deploying Voice AI in Production
Real-world infrastructure lessons from teams building production AI voice systems
Insights and ResourcesRunning Repeatable LLM Latency Benchmarks with Telnyx stateful actors
How Telnyx stateful actors cut TTFT by 67% for MiniMax M3 and 32% for GLM 5.2 in repeatable LLM inference latency benchmarks, and why stateful compute fits production AI workflows.
Conversational AIWhat Is Voice AI Latency? Typical Numbers, Standards, and How to Measure It
Learn about voice AI latency, benchmark response time, and how it is measured end to end for real phone calls.
- Conversational AI
How to Build an AI Audio Translator with Telnyx STT, Inference, and TTS
Learn how to build an AI audio translator with Telnyx STT, AI Inference, and TTS. Transcribe source audio, translate the transcript, and generate target-language speech in one Flask pipeline.
- Conversational AI
How to Build a RAG Application with Telnyx Inference
Learn how to build a retrieval-augmented generation (RAG) API with Telnyx AI Inference. Use Telnyx embeddings to vectorize a knowledge base, retrieve relevant documents, and send grounded context to Telnyx chat completions for accurate answers.
- Inference
How to Extract Structured JSON from Messy Text with Telnyx AI Inference
Use Telnyx AI Inference to convert messy text like support tickets and inbound emails into predictable JSON objects your app can route, store, and act on.
- Inference
How to Extract Structured JSON from Messy Text with Telnyx AI Inference
A practical example of using Telnyx AI Inference to convert unstructured text like support tickets and emails into clean, structured JSON your application can route, store, and act on.
InferenceGLM 5.2 Inference benchmarks across 4 providers
This report compares GLM 5.2 inference latency across Telnyx, Baseten, Together AI, and Fireworks using the same six prompt profiles and 10 streamed runs per provider/profile cell.
- Inference
MiniMax M3 Runs Best on Telnyx Inference
We benchmarked MiniMax M3 across Telnyx, Together AI, and Fireworks. Same model, same prompts, same streaming setup. Telnyx delivered the fastest E2E latency, 21% higher throughput, and the most consistent performance.
Consumer Insight PanelsTransparency drives Voice AI trust: 38% value disclosure most
Voice AI Consumer Insight Panel—December 2025
Conversational AIYour Voice AI Agent works until it doesn't. Here's why.
The hard part of building voice AI agents isn't connecting APIs. It's writing instructions that prevent the LLM from skipping steps, improvising, and breaking.
- Conversational AI
From Prompt to Phone Call: A Voice Developer’s Guide to the Telnyx AI Repo
Use the Telnyx AI repo as a practical workflow for building, provisioning, and iterating on voice agents with AI coding assistants.
- Inference
Inference Benchmark: Which Latency Metric Should You Optimize For?
A head-to-head latency benchmark of three leading inference providers across 540 streamed requests.
TelnyxAIDA: A pragmatic look at how we use AI Agents internally
Building an AI assistant that lives in Slack, remembers context with Mem0, and scales through multi-agent architecture.
