HNHacker News
TopNewBestAskShowJobs

mcemilg

143 karma · joined June 17, 2015

https://mcemilg.dev

https://github.com/mcemilg

https://twitter.com/mcemilg

submissionscomments
mcemilg··on Whisper.cpp example running fully in the browser
Thanks for your good wishes. It's not going well actually, the downloads are very few after 10 days and no paying customers. The App Store seems like a desert, even there are not many competitors in transcription, still the app impressions and page views are very few. What I see that people trying to do marketing outside the app store.
mcemilg··on Whisper.cpp example running fully in the browser
I recently laid off and currently trying to build some apps that could create some revenue that can afford my costs for some weeks. I built a transcription and dictation app for Mac [0] using whisper.cpp, small model works really well on 2019 mpb and m1 for streaming (dictation). It was really straight forward to use, however the streaming algorithm doesn't ready for production so I implement my own algorithm using VAD. I believe in that with that pace this could also be fixed.

[0] https://apple.co/3j2k8E7

mcemilg··on Sergey Brin: Irate Call from Steve Jobs
I am always wondering how these recruiter knows that "He is absolutely one of the best in the world at Browser technology" and how to be in this level in any area of expertise.
mcemilg··on Ask HN: Who wants to be hired? (November 2022)

  Location: Turkey
  Remote: Yes
  Willing to relocate: Within EU
  Technologies: Python, SQL, PyTorch, Keras, Pandas, NumPy, Scikit-learn, XGBoost, Docker, PostgreSQL, MongoDB, Grafana, Prometheus
  Résumé/CV: https://bit.ly/3zCklDj
  Email: mcg [at] mcemilg [dot] dev
  Github: https://github.com/mcemilg
Machine Learning Engineer with a master’s degree in Computer Engineering and 5+ years of experience to research, develop and deploy Deep Learning and Machine Learning applications. Worked with agile teams, collaborated with researchers and business people, and built products that gained customers and funds in less than a year. Recently laid off because of cost-cutting policies and actively seeking for similar positions.
mcemilg··on Ask HN: Who wants to be hired? (October 2022)

  Location: Turkey
  Remote: Yes
  Willing to relocate: Yes
  Technologies: Python, SQL, PyTorch, Pandas, NumPy, Tornado, Docker, Postgresql, MongoDB, Grafana
  Résumé/CV: upon request
  Email: mcg [at] mcemilg [dot] dev
Machine Learning Engineer with a master’s degree in Computer Engineering and 5+ years of experience to research, develop and deploy Deep Learning and Machine Learning applications. Headed small teams, collaborated with researchers and business people, and built products that gained customers and funds in less than a year.
mcemilg··on Hacker News “Who is Hiring?” top-level comments over time
From another point I would like to check who wants to be hired posts by comparing who is hiring. I can see that hundreds of ML engineers and Data Scientists looking for a job but on the other side there are not that many jobs for ML folks.
mcemilg··on Fast (2019)
It is nonsense to compare the timeline of those products in days. I think three main things are missing, budget, know-how, and how many people are involved which multiplies the working days.
mcemilg··on End-of-life dates
It's cool; honestly, I didn't know Python 3.6 support ended two months ago while I was still using it.
mcemilg··on Completing a part-time Master's in computer science while working
I've almost done this too but I saw that people having full time in masters or researcher in a lab more successful than people doing part-time masters as me. They have more time to do research. They produce more because it's their main job, they can go deeper. They collaborate more with researchers that also broader their research and network in the area. It can be the hardest thing to write a paper from research idea to result alone or with an advisor which is the advisor only will help you with the theory writing.
mcemilg··on Deep Learning Interviews book: Hundreds of fully solved job interview questions
The ML/DS positions highly competitive these days. I don't get why ML positions requires hard preparations for the interviews more than other CS positions while you do similar things. People expect you to know a lot of theory from statistics, probability, algorithms to linear algebra. I am ok with knowing basic of these topics which are the foundations of ML and DL. But I don't get to ask eigenvectors and challenging algorithm problems in an ML Engineering position at the same while you already proof yourself with a Masters Degree and enough professional experience. I am not defending my PhD there. We will just build some DL models, maybe we will read some DL papers and maybe try to implement some of those. The theory is the only 10% of the job, rest is engineering, data cleaning etc. Honestly I am looking for the soft way to get back to Software Engineering.
mcemilg··on Applebot
I was scared while writing crawlers to various sites since I don't know how far it's legal, but Apple says look, this is my crawler.
mcemilg··on Why GPT-3 Matters
Take gpt-2, bert or any other attention based language model. Apply few shot learning on any domain to model you take. You will not see so much meaningful difference that matters. Even the GPT-3 super large than the other models. There is a hype because people able to use that domain adaptation easily without serving huge ml models by themselves. OpenAI presents that serving already. That makes it easy to explore language models to developers that don't know ML that much to tune.
mcemilg··on Why GPT-3 Matters
We saw that the GPT-3 doesn't matter that much, right?
mcemilg··on OpenAI API
From AGI to money machine...
mcemilg··on Hosting a GPT-2 autoresponder bot
Serving an app to 1k daily active user cost you for just $300/month but for similar daily user with an nlp app will cost you $3000/month. I think the GPU vms are overpriced because of the AI hype. Buying a GPU server like the old days is much more cheap.
mcemilg··on Echo Frames – Eyeglasses with Alexa
Why not just headphones but glasses?
mcemilg··on YouTube should stop recommending garbage videos to users
How to define garbage?
mcemilg··on Show HN: A Mood Tracker on the Web
That's an awesome idea for this kind of applications. Does anybody know a sample app using this feature?
mcemilg··on Translatotron: An End-to-End Speech-to-Speech Translation Model
Same.
mcemilg··on Ask HN: Who is hiring? (May 2019)
Hi,

I am interested with ML Engineer. I want to ask some questions to you. I am not in US, I will need visa sponsorship. In the job posting it seems you can be sponsor to H-1B. Can you give some info about hiring process? Did you hire anyone with H-1B in the past? If everything are done successfully, how long does it take?

Thanks!

mcemilg··on Choose Firefox Now, or Later You Won't Get a Choice (2014)
I was very afraid about syncing my data with other devices on Firefox. But today I switched to Firefox and passed all my chrome data (even stored passwords) to Firefox. It can sync with all devices such as mobile. I hope I will not get trouble about web apps :P
mcemilg··on Christie’s sells its first AI portrait for $432k
This is a sad story. But Everybody needs to get some lesson from this case if they publish their work as open. The open source licenses needs to update for machine learning models.
mcemilg··on Ask HN: 4 months after posting my project on HN, it has 0 users. Has it failed?
I think you need to make the starter pack to free or create a very tiny pack for free starting. If the project really valuable to customer trust it they will buy more. But if it is not it didn't money anyway, no need to be sad. This will happen to anybody.
mcemilg··on Dear Spotify, please let me unlink my Facebook account
And also I cannot change my name if I will not add facebook account. It is very ridiclous. They are very huge startup but they did not much time for adding a simple form. :P
mcemilg··on Interactive workflows for C++ with Jupyter
Okay. If its works amazing, so it can be work on other compiled languages? This will be amazing and it can be game changer.
mcemilg··on Show HN: Nearly 50% of the tensorflow tagged questions on SO are unanswered
Yeah you are right. A momentary blindness.
mcemilg··on Show HN: Nearly 50% of the tensorflow tagged questions on SO are unanswered
Thanks for your interest, I found the answer from my project advisor. In this case I will answer my own question.

The problem you explained is the same to the other topics. But unanswered rate is not that high. I think the tensorflow question are hard to answer even it can find in internet. And also answering a question gets lots of time. In this case lots of people do not give this time for a question.

mcemilg··on Show HN: Nearly 50% of the tensorflow tagged questions on SO are unanswered
I've shared a question with tensorflow tag on stackoverflow. It was not answered and the number of views is very low. Then I noticed that, many of the questions asked before me share the same fate as my question. Many of them were unanswered. 19,225 questions were shared with the Tensorflow tag and 9,391 were unanswered.
mcemilg··on Science and Tech of the Byzantine Empire
Well yeah I saw it. So then we know at least there are three people exists in the history who they try to fly with robe. The Arab, Hezarfen and the French guy. People are weird. These are some kind of wingsuit prototype.
mcemilg··on Science and Tech of the Byzantine Empire
it is necessary to investigate their correctness. The "On fatal wings" story is similar to Hezarfen's story. Maybe it's a story that has been around for years.Take a look at that https://en.wikipedia.org/wiki/Hez%C3%A2rfen_Ahmed_%C3%87eleb...
← PreviousPage 2 of 3Next →