With more than 10 years of software development experience, I have developed a strong foundation in cyber-security, web development, machine learning, and cryptocurrency/blockchain. My background in Astronomy and Astrophysics (H. BSc and MSc) has provided me with strong analytical, mathematical, programming, and communication skills.
In the cyber-security domain, I have spent four years mastering Golang, PostgreSQL, Grafana, and GitHub Actions. At Arctic Wolf, I was a team lead for the past two years, where I managed five developers. I guided the team on a modularization project of our end-point Agent, and where I managed the biannual release cycle of our software. My back-end web development experience spans five years, where I have worked with Node.js/PHP/Golang, MySQL/PostgreSQL, Redis, RabbitMQ, and Docker/Docker-Compose. I also have more than a decade of experience in scientific programming, using FORTRAN, C/C++, Python and Bash.
Some of the work I am most proud of includes the parallelization of the SWIFT numerical integrator, a ML document classifier for OCR text, my contributions to the Bitcoin code base, and my entry in the 2016 Halite AI Challenge where I ranked in the top 3%. To see my ML portfolio, visit: huggingface.co/spaces/ccapo/portfolio
Outside of my professional career, I am passionate about pushing the boundaries of my personal projects. Currently, I am building a high-altitude balloon equipped with a Raspberry Pi and an Arduino. I also enjoy crafting rogue-like video games, blending creativity with interesting game mechanics. For a glimpse of my work, visit: github.com/ccapo
With 13 years of experience in software development, I have developed a strong foundation in cybersecurity, web development, machine learning, and cryptocurrency/blockchain. My background Astronomy and Astrophysics (H.BSc and MSc) has provided me with strong analytical, mathematical, programming, and communication skills.
In the cybersecurity domain, I have spent four years mastering Golang, PostgreSQL, Grafana, and GitHub Actions. My back-end web development experience spans five years, where I have worked with Node.js/PHP/Golang, MySQL/PostgreSQL, Redis, RabbitMQ, and Docker/Docker-Compose. I also bring over a decade of experience in scientific programming, utilizing languages such as FORTRAN 77/90, C/C++11/17, and BASH/Python.
When I came across the site years ago, I saved a copy of it so I could run it locally in my browser. Only later did I realize there was a GitHub repo: https://github.com/cantonbecker/bitcoinpaperwallet
The detector could just be the entire galaxy, it doesn't have to be an artificial construct.
To your question, you have already seen it, hiding in plain sight.
Lark Lane, Liverpool seems quite apposite here. Two Turkish restaurants, an Italian-American one, a Mediterranean one, a Thai one, three Chinese takeaways, a burger place, an Indian restaurant, a Tandoori house, a pizza place, a middle-eastern restaurant, a Greek restaurant, and two bistros.
One of the takeaways is a vegan "Chinese chippy" named Woo Tan Scran.
I remember reading about Kahan summation, and I naively thought it would help improve the accuracy/consistency of a n-body code I was using for my PhD. I was trying to resolve a discrepancy between summations when performed serially versus in parallel using OpenMP. In the end I abandoned the approach, since it seems the issue lay elsewhere.
I've employed this Kahan summation in a different physics simulation that was giving parallel vs serial discrepancies due to global sums that sum in different orders in parallel. It did help, but I ultimately concluded that it isn't a complete solution for this issue... in short, I think it just bumps the problem several orders of magnitude down.
For what it's worth, and in the event that your issue was indeed the differing order of summing-- there is a solution that /does/ work, which is to cast the numbers to a large (eg, 256 or 512 bit) fixed point representation and leverage the reproducibility of integer sums.
The issue here is that unless you have checks that your values dont over or underflow at every stage of the calculation, you can never actually know if your answers are accurate, even if they might be repeatable.
I would think that in most cases you should be able to prove an upper bound on the sum and make sure the accumulator is large enough to accommodate it. My guess is that 512 bits should be enough to span all physically/computationally relevant scales.
Note that digits of precision are expressed in decimal unless explicitly given as “bits”, so that the phrase “10,000-digit arithmetic” really means around 33,000 bits.
https://www.jobsearchdb.com
reply