init

    Location: Sweden, EU

    Remote: Yes (UTC +/- 5)

    Willing to relocate: No

    Technologies: Go, Python, Java, C++, GCP, AWS, Terraform, Docker, Linux, CI/CD, WebRTC, gRPC, PostgreSQL, BigQuery, PyTorch, LLM APIs

    Résumé/CV: Available on request

    Email: (initproc) at (proton.me)

Looking for fractional, advisory or short-time consulting engagements.

10+ years building infrastructure and backend systems at FAANG, FAANG-adjacents and early-stage startups. I've led projects across the full stack, from cloud infrastructure to product, through developer productivity and ML/AI pipelines, payments and real-time communication systems.

I'm particularly useful early on: greenfield architecture, getting a team to move fast without accumulating debt, and navigating compliance (SOC2, GDPR) without it becoming a distraction.

I've built and worked on this exact problem before at bigtech, startup and personal projects.

Regex works well if you have a very limited set of sender and recipient accounts that don't change often

Bayesian or DNN classifiers work well when you have labeled data.

LLMs work well when you have a lot of data from lots of accounts.

You can even combine these approaches for higher accuracy

  Location: Sweden, EU
  Remote: Yes (can work within UTC +/- 5)
  Willing to relocate: No
  Technologies: C++, Java, Python, Go, Linux, GCP, AWS, Terraform, Bazel, CI/CD, Github Actions, WebRTC, Data processing (big and small), Machine Learning and AI (Tensorflow, LLM APIs and local models).
  Résumé/CV: Available on request
  Email: (initproc) at (proton.me)

Looking for short-term or part-time consulting projects.

Background: 10+ years experience across startups and big tech (ex Google and Stripe). Lead and contributed to projects in compliance/security, distributed systems (Fintech, payments processing, real-time communication), machine learning and AI, as well as developer productivity and release engineering.

6700 / 10000. I guess because I didn't zoom on the map and got the time off by 34 years of the actual event.

The scoring could be improved.

This reminds me of Microsoft FrontPage more than 20 years ago.

> Can you explain in what way did this manifest in Sweden for instance? Or have any data/evidence besides some semi-vague claims about antisemitism and some immutable characteristics of European societies (which are hardly monolithic to begin with).

If you live in Sweden then you will also know that the state puts refugees in the same areas (Rinkeby, Vivalla, Tensta, etc ...). These areas are then labeled as unsafe because of a slightly elevated crime rate and because they're labeled unsafe, swedes start moving out and quality of services and house prices drop and the downward spiral continues until the area becomes a ghetto even though they're usually not that bad.

Although SFI exists to teach Swedish to immigrants, the quality of the teaching is not great in most schools.

That's where the integration effort stops.

Even professionals who move to Sweden for work have a hard time integrating in Swedish society. That's how you end up with people living in upscale parts of Stockholm for 10+ years and can barely form a sentence in Sweden.

Making their APIs easy to use like they used to be 10 years ago will be equivalent to releasing their core assets. In the past you could do almost anything from the Facebook API that you could do with their web or mobile app.

They release a lot of open source stuff as other commenters have mentioned but you can't build a Facebook or Instagram competitor just by integrating those components.

In a past job I built a code search clone on top of Kythe, Zoekt and LSP (for languages that didn't have bazel integration). I got help from another colleague to make the UI based on Monaco. We create a demo that many people loved but we didn't productionize it for a few reasons (it was an unfunded hackathon project and the company was considering another solution when they already had Livegrep)

Producing the Kythe graph from the bazel artifacts was the most expensive part.

Working with Kythe is also not easy as there is no documentation on how to run it at scale.

I've used both Code Search and Livegrep. No, Livegrep does not even come close to what Code Search can do.

Sourcegraph is the closest thing I know of.

Henry Kissinger has a heavy German accent that didn't prevent him from becoming one of the most influential American politicians of the 20th century.

I worked on something like this more than a decade ago back when bayesian classifiers were SOTA for sentiment analysis. It's relatively easy to use an existing model to bootstrap and fine tune a new model. The hardest part is collecting and cleaning up the data for training.

Unfortunately this has already been happening for a decade or more since the NATO "intervention" in Libya and it has little to do with climate change.

I'm interested in this space and other mining-adjacent industries in the DRC-Zambia copper belt region due to the potential of this region to be a key player in central Africa's development and the world's move into electric vehicles.

The main challenge is getting capital.

Let's get in touch!

Thank you for sharing this link. The Internet Archive is truly a marvel of unequaled value. Didn't know it had such a trove of warez without the risk of pwnage that warez and torrent sites came with.