HN user

wmantly

27 karma
Posts0
Comments15
View on HN
No posts found.

Location: New York City

Remote: Yes (Remote or hybrid/in-person in NYC)

Willing to relocate: No

Technologies: Python, JavaScript, Node.js, Go, PostgreSQL, ZFS, Linux (Mint/Debian/Ubuntu), Proxmox VE, Docker, Git, AWS, GCP, OpenAPI, MCP (Model Context Protocol), local LLM deployment (Ollama, vLLM), agentic RAG infrastructure.

Résumé/CV: https://william.mantly.vip (At bottom of page)

Email: william@mantly.vip

Hey HN,

I’m a senior engineer and infrastructure architect who has spent the last 20+ years building everything from high-concurrency data pipelines to bare-metal server clusters. I’ve worn the corporate VP hat at JPMorgan Chase leading teams managing over 100,000 servers, and I’ve worn the founder hat running my own consulting firm, Theta 42.

But honestly? I miss the early stage startup energy. I miss the environments where we pivot fast, build from a whiteboard to production, and where everyone dons multiple hats to smash the next milestone.

What I bring to your team:

Real-world leadership & mentorship: I don't just manage tasks; I grow engineers. I've led development teams from scratch, run a tech consulting firm, and even spent years as a head instructor teaching Python/Django backends—maintaining a documented 90% job placement rate for my graduates. I know how to unblock a team and keep execution high.

Deep infrastructure, backend roots: I'm a self taught engineer who started with Perl in '99 on a WebTV. I’m a native Linux user who deeply understands networking, hypervisors, and storage layers (think Proxmox, LXC, and ZFS), not just cloud abstractions.

Pragmatic AI execution: I’m actively building local AI pipelines, structured tool-use interfaces, and agentic RAG infrastructure. I know how to cut through the LLM hype and implement local models (Ollama/vLLM) that actually solve business problems.

If you’re an early-stage startup or a tightknit team looking for a seasoned technical leader who can anchor your backend, architect your infrastructure, and mentor your engineers without losing the hunger to get into the weeds and write code, let’s talk.

Me too, I used 10.04(i think), the last gnome version until the repo stopped updating. Linux Mint has been my desktop of choice ever since. With their MATE and Cinnamon version looking like a DESKTOP and everything works out of the box its win/win. In 2010~ every one lost their collective mind when it comes to UI.

I have been using Linux on many higher end HP laptops without issue. At the moment I have HP Spectre x360 15in 2016,8th gen i7 16gb RAM, 512GB SSD and 2gb mx950 Nvidia. After installing Linux Mint 19 everything worked out of the box and i get about 6-hour of battery. I have a few older HP envy series and each of them work with out issue. The only issue i have had in the past 5years is laptops with more then 2 speakers. Linux just assumes 2 speakers and you get shit sound. This is fixable with jack-reset.

This is from a 4 socket Xeon E7-4860 with 64 ram slots(16 in use)

  e7-4860:~ Mon Jun 11
  03:06 PM william$ numactl --hardware
  available: 4 nodes (0-3)
  node 0 cpus: 0 1 2 3 4 5 6 7 8 9 40 41 42 43 44 45 46 47 48 49
  node 0 size: 16035 MB
  node 0 free: 1306 MB
  node 1 cpus: 10 11 12 13 14 15 16 17 18 19 50 51 52 53 54 55 56 57 58 59
  node 1 size: 16125 MB
  node 1 free: 3237 MB
  node 2 cpus: 20 21 22 23 24 25 26 27 28 29 60 61 62 63 64 65 66 67 68 69
  node 2 size: 16125 MB
  node 2 free: 11004 MB
  node 3 cpus: 30 31 32 33 34 35 36 37 38 39 70 71 72 73 74 75 76 77 78 79
  node 3 size: 16123 MB
  node 3 free: 12044 MB
  node distances:
  node   0   1   2   3 
    0:  10  20  20  20 
    1:  20  10  20  20 
    2:  20  20  10  20 
    3:  20  20  20  10
The chart at the bottom of the output is the weight for accessing a memory pool from a CPU socket. This is the most important part of the output.

On this server, CPU socket 0 is hardwired to ram slots 0-15

CPU 1 to ram slots 16-31

CPU 2 to ram slots 32-47

CPU 3 to ram slots 48-63

If CPU 0 wanted to read something outside of its local ram slots, it would have execute something on CPU n, then copy that segment to its local ram group.

Where I work, if the smoke break isn't solo, it's a meeting. We even take non-smokers that would be relevant to the conversation. Sometimes us smokers are the only single to non-smokers that maybe you should get up for a few minutes. I use to work in windows building, after my first few days my co-worker who doesn't smoke thanked me for being smoker, he worked there for about a year prior and never knew when he should get up and go outside, I gave him that guidance.

Every job I have worked that isn't hourly, break times are not monitored. If you have something during the work day, just go, if you want to walk around the park and clear your mind, have fun, need a smoke, shit have 2. So long your not screwing something up timing wise.

Also, it seems to me that the people who complain about such things have unimportant roles...

This is from using Debian stable, and normal. Debian has 4 "branches"

1. Old stable, the previous stable version and only receives security updates.

2. Stable, the current stable version. There is a massive amount of stability testing and security hardening that go into ever package before its allowed in this branch. This is usually a very long process and the packages are usually old.

3. Testing, This is where packages live while they are being tested for stable. This branch may be broken, but is generally working during the release cycle, and may not receive security updates in timely manner. 4. Un-stable, This is bleeding and mostly in broken state...

During the release cycle, packages are selected and brought closer and closer to 'stable' and move though the branches. At a certain time, the package version are frozen in a feature freeze, and are not allowed to have feature updates, only stability and security hardening. When are the packages are ready, the debian team releases alpha and beta builds to test everything. Then, and only then is a new stable released.

If every thing had the Q&A of debian, nothing would crash, but everything would be old.

This is the reason why so many people use one of the many distro based up of debian testing. Ubuntu, Linux Mint and many more are based off of testing, with newer packages and there own feature freeze. The ubuntu cycle tries work like clock work every 6 months and debian has more of "when its ready" mind set.