braindead_in

Here's a free transcript, in case anyone's interested (around 80% accurate).

https://scribie.com/account/eeIvFBUv?fid=ff499dd6f79e41b682c...

Is there any editor which mimics the look and feel of Google Docs or MSWord?

Scribie | Deep Learning Engineer | Bangalore, India | Full time, one-site

We build tools and software which reduce the manual effort required for high accuracy audio/video transcription by 80%. We are looking for a Deep Learning Engineer with hands on experience in the following:

  - RNNs/CNNs/LSTMs 
  - Language Modeling using nGrams/RNNLMs etc. 
  - Seq2Seq networks 
  - TensorFlow or PyTroch

Nice to have's:

  - A Masters/PhD 
  - 3+ years of experience 
  - Ambition to become the Head of R&D 
  - Ability to work independently without supervision 
  - High levels of self-motivation

Apply at https://angel.co/scribie/jobs/240398-deep-learning-enginner

Scribie | Full Stack Developer | Bangalore, India | Full time, one-site

We are building an AI for audio/video transcription and looking for a full stack developer with experience in the following skills.

Must have's: Node.js, PHP, HTML/CSS/JS, MySQL, Python, Shell scripts, AWS hands-on experience

Good to have: A bachelor's degree, 3-5 years of experience, comfortable with Agile methodology, believer in TDD, love technical challenges, ability to work independently without supervision, high levels of self-motivation

Apply on https://angel.co/scribie/jobs/235433-full-stack-developer

Scribie | Full Stack Developer | Bangalore, India | Full time, one-site

We are building an AI for audio/video transcription and looking for a full stack developer with experience in the following skills.

Must have's: Node.js, PHP, HTML/CSS/JS, MySQL, Python, Shell scripts, AWS hands-on experience

Good to have: A bachelor's degree, 3-5 years of experience, comfortable with Agile methodology, believer in TDD, love technical challenges, ability to work independently without supervision, high levels of self-motivation

Apply on https://angel.co/scribie/jobs/235433-full-stack-developer

Any WER benchmarks for TED, Librisvox, etc?

Nice! Would you be supporting Volta GPUs when it is launched?

Most datasets are 16khz sampling rate, wav file. A good place to start is Mozilla deepspeech implementation.

What about hand drawn figures?

What about PSDs? PSD to html is a big market and a product like this could be a good fit.

Predicting mistakes made by humans during audio/video transcription, giving them auto-complete suggestions, classifying severity of changes, predicting the difficulty level based on audio characteristics.

So when are the new AWS instances are coming?

I remember seeing a paper from Microsoft on arXiv about the recent advancements in ASR which had a WER graph. Can't seem to find it now.

I definitely plan to try it soon.

1. Clone DGX-1

2. Rent it out.

3. Profit!!!

I would pay for cloud DGX-1 servers.

Why not use DGX-1? Surely Google can afford a bunch of these.

Thanks for the explanation. Will it work if there are gaps in the transcript? Eg, the clean verbatim transcript where the ah's and uhm's are left out.

What's the accuracy level of alignment?

Are there any examples of seq2seq networks being used for tasks other than NLP? For example, can it be used for something like noise removal?

How accurate is the speaker diarization? How does it compare to LUIM?

Mozilla's DeepSpeech has a native client https://github.com/mozilla/DeepSpeech

No search?

Scribie | Data Scientist | San Francisco/Bangalore | ONSITE

Scribie is an human powered audio/video transcription service and we are looking to build an speech recognition system using our own data.

You must have hands-on experience with the following:

- TensorFlow

- Acoustic Modeling

- Language modeling

- Deploying models in production.

Drop me a line at rajiv@scribie.com if you're interested.

VM migration between the cloud hosting services is a big pain. I hope someone comes up with an app to do that.

Or maybe treat the disease? No?

If history has taught us anything, it is that the we survive. Wars, disease, plague has not killed us off. We will survive this as well. The question is at what cost. I am not ready to give up my humanity. That is the only thing separating me from a monkey.

And US is not the only country affected by Islamic terrorism. A visa ban is hardly the magic panacea for terrorism.

I question how effective this policy is going to be in stopping the actual terrorists. Just as a determined hacker will always find a way to break into your server/network, the terrorists will find a way to illegally enter.

No standalone ASR? I am looking for an ASR as a Service for audio/video files where I can use my data to train my own models. AFAIK nobody's offering that yet.

SEEKING FREELANCER, SFO, Remote Ok

Scribie, Speech Recognition Engineer

Scribie is an audio/video transcription service where we manually convert interviews, meetings, teleconferences, podcasts and other spoken audio files to text via our 4-step process which guarantees high accuracy. We have tons of high quality data, audio files and their corresponding transcripts and we are looking for someone who can build acoustic and language models using that data.

The key skills we are looking for are:

- Hands-on experience with speech recognitions systems, eg. Kaldi

- Knowledge of major components of an ASR systems

Drop me a line at rajiv@scribie.com if you're interested.

https://scribie.com Email is in my profile as well.

We are an human powered audio/video transcription service and have lots of training data that can be used for training a speech recognition system. An ASR-as-a-Service in the cloud kind of platform where we can use our data to continuously train and improve the models would be very useful for us.