Skip to content
View sachaarbonel's full-sized avatar
πŸ‘¨β€πŸ’»
Uncovering bugs
πŸ‘¨β€πŸ’»
Uncovering bugs

Organizations

@GetStream

Block or report sachaarbonel

Block user

Prevent this user from interacting with your repositories and sending you notifications. Learn more about blocking users.

You must be logged in to block users.

Please don't include any personal information such as legal names or email addresses. Maximum 100 characters, markdown supported. This note will be visible to only you.
Report abuse

Contact GitHub support about this user’s behavior. Learn more about reporting abuse.

Report abuse
sachaarbonel/README.md

Hi, I'm Sacha πŸ‘‹

Twitter β€’ LinkedIn β€’ GitHub β€’ Medium β€’ Email

I’m a Software Engineer with 7+ years in backend and mobile. I’ve shipped systems that reached over a billion users and ran services that process millions of messages a day. I care about open source, clear design, and reliable scale.


What I work on

  • Pion/WebRTC–based media I/O library used by other services (egress/ingestion for transcription & recording; ingress for HLS, WHIP, SRT).
  • Whisper infrastructure: language hot-reload config (no restart), Prometheus/CloudWatch metrics, Grafana dashboards, rollout speed-ups with Packer-baked AMIs, and CloudFormation maintenance.
  • VAD with Silero via ONNX Runtime (Go) to enable real-time captions (no VAD β†’ no live captions) and reduce waste.
  • Accuracy work: built a hallucination dataset (transcribed a noise corpus) + Aho-Corasick matcher in Go, with deloops and Unicode range filters to remove hallucinations; important for most customers and critical for healthcare.
  • Codegen & SDKs: internal OpenAPI β†’ server-side SDK tooling; owner of the Python and Go SDKs (35k+ LOC each).
  • Product support: outages on rotation, post-mortems, indexes for big imports, quick fixes across chat and dashboard (picked up Django fast when needed).

Highlights

  • p95 live transcription/closed captions: ~650 ms β†’ ~300 ms.
  • Cost: ~36Γ— cheaper than OpenAI 4o transcribe for our load.
  • Rollouts: ~11 min β†’ ~1 min with Packer AMIs + trimmed Puppet.
  • One service for transcription and captions β†’ ~50% infra/ops reduction.
  • Proved CPU wasn’t the bottleneck; eased GPU concurrency with NVIDIA MPS.
  • Quality: sales team dogfoods our app as a Gong replacement.

Tech

Go β€’ Rust β€’ Python β€’ C++/CUDA β€’ Whisper β€’ ONNX Runtime β€’ Pion/WebRTC β€’ AWS β€’ CloudFormation (maintenance) β€’ CloudWatch β€’ Prometheus β€’ Grafana β€’ Packer/AMIs β€’ Puppet β€’ Kibana β€’ Nsight β€’ NVIDIA MPS β€’ OpenAPI

Writing

Contact


Sacha's GitHub stats

Pinned Loading

  1. huggingface/transformers huggingface/transformers Public

    πŸ€— Transformers: the model-definition framework for state-of-the-art machine learning models in text, vision, audio, and multimodal models, for both inference and training.

    Python 149k 30.2k

  2. ggml-org/whisper.cpp ggml-org/whisper.cpp Public

    Port of OpenAI's Whisper model in C/C++

    C++ 42.7k 4.6k

  3. huggingface/candle huggingface/candle Public

    Minimalist ML framework for Rust

    Rust 17.9k 1.2k

  4. pgcentralfoundation/pgrx pgcentralfoundation/pgrx Public

    Build Postgres Extensions with Rust!

    Rust 4.1k 290

  5. guillaume-be/rust-bert guillaume-be/rust-bert Public

    Rust native ready-to-use NLP pipelines and transformer-based models (BERT, DistilBERT, GPT2,...)

    Rust 2.9k 236

  6. reefdb reefdb Public

    ReefDB is a minimalistic, in-memory and on-disk database management system written in Rust, implementing basic SQL query capabilities and full-text search.

    Rust 89 3