HN user

shrike

870 karma

Craig Carl

craig.carl.ca@gmail.com

Posts11
Comments107
View on HN

Location: San Francisco Bay Area

Remote: OK

Willing to relocate: No

Technologies:

  ** Full-stack web (Postgres, Express, React, Node), Cloud infrastructure (AWS, GCP, OCI)
  ** PyTorch, Apache Spark, Apache Airflow, Kafka, Apache Aerospike, Apache Flink
  ** Postgres, MongoDB, Redis/Dragonfly, Cassandra
  ** CephFS, GlusterFS, ZFS, scaling and geo-distributing filesystems, LTO/LTFS
  ** LLM Integrations: Claude, OpenAI, Gemini, Mistral, DeepSeek, routers
  ** LLM fine-tuning, RAG/CAG, LoRA, MCP, Agentic workflows, GenAI, deep integrations
  ** Cloud Automation: Terraform, AWS Systems Manager, OCI Terraform
  ** Scaling, Clustering, Optimizing heavy throughput systems
Recent History:
  ** SOC and ISO pre-audit reviews for startups in the Storage as a Service, Healthcare, and Finance spaces.
  **  AI Startup Focused on Legal Discovery Management. Exited via acquisition.
Résumé/CV: https://drive.google.com/file/d/1YX3xporepZBDFq_mJQ062u6YHNN...

Email: craig.carl.ca@gmail.com

     Location: San Francisco Bay Area
     Remote: OK
     Willing to relocate: No
     Technologies:
        * Full-stack web (Postgres, Express, React, Node), Cloud infrastructure (AWS, GCP, OCI)
        * PyTorch, Apache Spark, Apache Airflow, Kafka, Apache Aerospike, Apache Flink
        * Postgres, MongoDB, Redis/Dragonfly, Cassandra
        * CephFS, GlusterFS, ZFS, scaling and geo-distributing filesystems, LTO/LTFS
        * LLM Integrations: Claude, OpenAI, Gemini, Mistral, DeepSeek, routers
        * LLM fine-tuning, RAG/CAG, LoRA, MCP, Agentic workflows, GenAI, deep integrations
        * Cloud Automation: Terraform, AWS Systems Manager, OCI Terraform
        * Scaling, Clustering, Optimizing heavy throughput systems
     Recent History:
        * SOC and ISO pre-audit reviews for startups in the Storage as a Service, Healthcare, and Finance spaces.
        * AI Startup Focused on Legal Discovery Management. Exited via acquisition. 
     Résumé/CV: https://drive.google.com/file/d/1YX3xporepZBDFq_mJQ062u6YHNNPz94V/view?usp=drive_link
     Email: craig.carl.ca@gmail.com

Location: Bay Area, CA, USA

Remote: Yes

Willing to relocate: Yes

Technologies: Hyper-scale IaaS and SaaS Infrastructure

Resume: https://www.dropbox.com/s/jlssjgy8x3amm9s/Craig%20Carl%20CV....

Email: hn@gestas.net

Relevant previous employers: AWS, Oracle Cloud (OCI), Splunk

Engineering and product leader focused on building effective teams that deliver solutions to delight customers. As a leader I foster a culture of customer obsession and ownership with a strong bias for action. I build strong engineering teams that deliver excellent product on-schedule. Experience hiring and developing globally distributed teams. Successful in reducing delivery costs and increasing direct and drag-along revenue. Track record of leveraging internal and external customer feedback to iterate on and improve product. Inventor with 7 issued patents.

Location: Bay Area, CA, USA

Remote: Yes

Willing to relocate: Yes

Technologies: Hyper-scale IaaS and SaaS Infrastructure

Resume: https://www.dropbox.com/s/jlssjgy8x3amm9s/Craig%20Carl%20CV....

Email: hn@gestas.net

Relevant previous employers: AWS, Oracle Cloud (OCI), Splunk

Engineering and product leader focused on building effective teams that deliver solutions to delight customers. As a leader I foster a culture of customer obsession and ownership with a strong bias for action. I build strong engineering teams that deliver excellent product on-schedule. Experience hiring and developing globally distributed teams. Successful in reducing delivery costs and increasing direct and drag-along revenue. Track record of leveraging internal and external customer feedback to iterate on and improve product. Inventor with 7 issued patents.

While that may be true, it's not what Spirit has said. They have reported that "...a non-standard manufacturing process..." was used to assemble "two of eight points where the vertical fin is attached to the fuselage.".

As an example; this could be as minor as using a riveting tool that hadn't been inspected on the correct schedule but was still operating as expected. It could be far more serious; perhaps that riveting tool was operating at a dangerously low pressure resulting in a poorly attached vertical fin.

We don't know yet. If this had happened before the 737 Max debacle I would have complete trust in Boeing's and the FAA's assessment that "...that there is no immediate safety issue.”. As it is I can't help being a little skeptical, that loss of trust is a real shame.

Starbucks did the same thing, arguing that -

“Wages and benefits are mandatory subjects of the collective bargaining process,” the company said. It rejects the union’s argument that it could offer the wage and benefit enhancements to unionized stores at any time.

The NLRB countered -

"...the NLRB said Starbucks violated labor law by offering raises and benefits — including increased training, career development opportunities, expanded tipping and even looser dress code policies — only to nonunion stores."*

Source: https://www.seattletimes.com/business/starbucks-shortchanged..., @ https://archive.ph/CrZTh

Wow, they sure do. Page 29 of the user guide for the Nespresso Zenius says -

This coffee machine is equipped with M2M (Machine to Machine) techology which may be activated in due time with your agreement. Thanks to a SIM card already integrated in the machine, such network connections will offer new services (subject to further terms and conditions) to its customers and improve the after sales process by automatically communicating machine troubleshooting / diagnostics to our Customer Relationship Centre (depending on country requirements and specificities).

https://www.nespresso.com/shared_res/manuals/zenius/www_Zeni...

AWS Tape Gateway 3 years ago

The VTL was invented because, at the time (2012ish), none of the largest enterprise backup solutions had a good S3 interface and none supported Glacier. After talking with the backup software vendors (Spectrum, Tivoli, Symantec, Commvault, etc) it became clear that adding another backup target wasn't something we (AWS) could get them to prioritize, for perfectly reasonable reasons. We could (and did) apply pressure via our shared customers, even then they estimated it would take years.

The fastest way to enable large enterprise access to S3 and Glacier for backups was to meet them where they were. We did this by virtualizing a tape library.

Background: I'm one of the original inventors - https://image-ppubs.uspto.gov/dirsearch-public/print/downloa.... I am no longer with AWS.

NOTE: The original article title, Florida Parents Partner with IJ to Shut Down Dystopian “Predictive Policing” Program is too long to submit. I took a substring, please adjust as appropriate.

Flexures are generally interesting, these flexures are amazing.

The video references the public documentation @ https://www.esmats.eu/amspapers/pastpapers/pdfs/2006/warden..... They weigh 665 grams (1.47lbs), have a range of 21mm (.826 inches) and a 7.7 nanometre (0.00000030315 inch) step. They work at temperatures down to -253C (-423.6F).

The cryocooler and these flexures have me seriously considering going back to school for a mechanical engineering/materials science degree.

I found this - https://www.bullionstar.com/blogs/jp-koning/golds-official-p.... The whole article is worth a read, this is the important bit -

“And that is probably why the U.S.’s statutory price of gold stays fixed at a decades-old level of $42.22. The consensus that independent central banking is a good thing (because it keeps a lid on inflation) dictates that the Fed have plenty of ammo. If the official gold price stays at $42.22, the Fed can lay claim to the full 261,498,927 ounces held by the Treasury. If the price is increased, the Fed gets only a sliver of that, the Treasury laying claim to the rest. And with fewer resources, the Fed’s has less control over the purchasing power of currency.”

I don't know if these are weird or not, I managed to lose ~25% of my weight (still lots to go) by -

1. Eliminating liquid calories and substitute sweeteners. Once I figured out that substitute sweeteners made me crave genuine sugars I got rid of them which made it easier to get rid of all liquid calories. Other than one or two servings of alcohol a month in social settings and an espresso in the morning I've had nothing but water to drink for a couple of years. It's made a huge difference in the way I feel.

2. I got rid of "direct" sugars. No ice cream, desserts, other sweets. I don't worry about carbs (bread, pasta, etc) being converted to sugar.

3. As a general rule I'm not hungry until noonish. If I forced myself to eat breakfast it triggered something and I ended up being hungry all day no matter how much I ate. I've decided the 3 square meals a day thing doesn't work for me. I switched to eating on an unstructured schedule, mostly salty snacks (chips, jerky, cheese, etc) during the day with a protein heavy meal at the end of the day.

About a month after I had ^^^ figured out I started feeling much more energized which meant I was able to do a lot more walking. I live in a city, 90% of everywhere I need to be on a regular basis is within a dozen blocks. Where a couple of years ago I might have driven 4 blocks now I walk everywhere.

The Institute for Justice is doing great work in this space. In addition to their civil forfeiture efforts I'm a huge proponent of their "License to Work" program and its efforts to limit onerous occupational licensing requirements. https://ij.org/report/license-to-work-2/. I especially like their approach to solving these problems. They defend individuals, this has an immediate and direct benefit. Then they use these cases to attract attention to the larger issue and drive regulatory change that way.

They are excellent stewards of donations, https://www.charitynavigator.org/ein/521744337. If you are looking for a good cause for your donations this is one. https://ij.org/support/give-now/

I disagree with their position on school choice, I still donate.

When my phone rings I want to know, with a high degree of confidence, who is calling me.

The only way I can think to do this well is by giving everyone a different phone number, like many of us do with email addresses. Implementing this would require a way to create and vend phone numbers quickly, similar to creating a password with a password manager. There would need to be a way to associate each number with who I gave it to. When my phone rang that identity would need to be surfaced in some way. I would need a way to make outbound calls using these numbers.

I'd happily pay $100+/month for <some number of> phone numbers (50-100?) plus $x/month for each number above that. If the service was metered by call minute there would need to be a generous number included in the monthly flat rate and excess minutes should be inexpensive.

I imagine the user experience would be similar to what Sidecar/Flyp/Cloud SIM deliver.

Absolutely they are. Nordstrom and Home Depot got called out for doing it 8 years ago - https://www.forbes.com/sites/petercohan/2013/05/09/how-nords.... The technology has progressed significantly since then.

I've worked on systems designed to send in-store customers custom, extremely time limited (< 1 hour) coupons based on where they are in the store and where they linger. Linger time being used as an indicator of purchase intent for a specific item. This does require that the customer has the store's app installed and running.

If you have access to an extended leave program try and take advantage. FMLA in the USA allows for 90 days of leave, my last employers disability insurance paid 80% of salary during leave. I expect there is something similar in the UK. All it took was a letter from a doctor, most of the people taking leave had stress related issues. The letters were primarily from psychologists. Several people submitted their notice in the last few weeks of leave.

Good reference - https://processormatch.intel.com/Resources/Program%20for%20C...

The regulations seem reasonable. They are limited to regulating the amount of power used while desktop systems are at idle, not while "in use". They take into account the potential for accessories to be added by the end user, the "Expandability Score". This means "bigger" computers can consume more power at idle. The CEC defines 4 idle states; short-idle, long-idle, sleep and off. Each state is well defined -

* Short Idle is system in S0 state, display on, no applications loaded or running, after a fresh re-boot.

* Long Idle is the same as Short Idle except the screen is off.

* Sleep is Power State S3 (or equivalent if the system doesn't directly implement S3).

The regulations also specify the required default timeouts for transitioning between idle states, e.g. the screen should be turned off after 10 minutes without user input, it should enter S3 after 15, etc. These don't apply if the system ships without an OS installed, they explicitly say that an end-user is free to change the defaults.

My first IT job was in a large call center. I was sweeping up in a data center and there was a keyboard cable stretched across a walkway. The keyboard wasn't movable (I don't remember why) so I unplugged it from the PC, swept around it, and plugged it back in. About 2 minutes later half a dozen people run into the room. Apparently the SUN workstation I had unplugged the keyboard from was a critical component of the call manager and there was a bug that forced a reboot when the keyboard was plugged in while the system was running. I had hung up on ~36,000 people. Lesson learned and I only had to keep the data center clean for another year.

My first IT job was in a large call center. I was sweeping up in a data center and there was a keyboard cable stretched across a walkway. The keyboard wasn't movable (I don't remember why) so I unplugged it from the PC, swept around it, and plugged it back in. About 2 minutes later half a dozen people run into the room. Apparently the SUN workstation I had unplugged the keyboard from was a critical component of the call manager and there was a bug that forced a reboot when the keyboard was plugged in while the system was running.

I had hung up on ~36,000 people. Lesson learned and I only had to keep the data center clean for another year.