sudoshred

Petition to have the question added to leetcode. Two variants one with a single type of parens such as () and one with mixed parens (),{},[]. In each problem implement the map method which takes a chunk sequence of the overall input, the reduce method which takes an array of map method output for each chunk and returns the overall answer.

Also consider RTS elements, like assigning certain schedulers to handle workloads which are swappable and tuneable, maybe with some skill tree unlocks.

Same people complaining "no one wants to work these days".

Depends. I can certainly offer to rewrite my old code that contributed to the memory shortage, for a price.

Safety of their IPO

Location: DC metro/Northern Virginia

Remote: Yes

Willing to relocate: Yes (continental US)

Technologies: TypeScript/Node, Python, Java, AWS, GCP, vector embeddings, ECS, SQS, DynamoDB, MongoDB, Datadog, CDK, Terraform, Docker, Kubernetes

Resume: https://drive.proton.me/urls/XKGXBP4BBW#qxa0grxxCT8c

Email: [email protected]

Senior Engineer most recently at Anywhere Real Estate (contacts platform, thousands of agents): designed and shipped duplicate contact detection using vector embeddings and weighted similarity scoring, replacing brittle boolean rules with fuzzy/near-duplicate matching that actually works on messy CRM data. Led a full redesign of async sync infrastructure—moved from Lambda to ECS, introduced a queue-backed worker pool (SQS + dedicated containers), and decoupled task execution via a shared worker contract. Took P99 enqueue-to-dequeue from ~5s to <10ms; cut peak queue ingress ~100× with producer-side backpressure. These patterns map directly to agentic workflows and human-in-the-loop automation at scale.

Before that: region build orchestration at Amazon (correctness invariants, async message passing, safer cutovers); ML-powered OCR and machine translation pipelines at The Bison Group (GCP Document AI + Azure Translator, scaled to 2,000-page docs, resolved >70 RPS throttling, zero-downtime API migrations); FHIR validation on Lambda at Cigna.

Comfortable owning ambiguous backend/AI feature work end to end—architecture, implementation, observability, rollout. I use AI tooling (Cursor, Claude, ChatGPT) heavily for design spikes, API scaffolding, and test generation, with review and monitoring before anything hits production.

Looking for: a team building something with real technical depth, where async/distributed systems or AI-enabled product features are central, not bolted on. Startup or growth stage preferred but not required. Open to fully remote or onsite/hybrid anywhere in the continental US—prefer to start remotely and happy to discuss relocation once we've established a good fit.

Statistically impossible to inadvertently generate a collision using UUID keys. UUID is designed to be unique when generated across any computer system. Practically speaking if you have an exactly matching pair of UUIDs from disparate system you have found the exact record match. The name gives a hint "Universally unique identifier". -Not a cryptographer.

sounds like a compliment to the implementation

Much like most office jobs

imminent danger pay

nearly retired

bitshifting screen renders directly to pixels on custom hardware, if I had to bet

Some people enjoy vendor locked managed services for their core infrastructure. Typically this decision is made when building from zero to one in resource constrained environments, and the long term play is to move to your own table/db when it becomes sustainable to do so. The only reason to move to a managed service after having done the work to setup self owned systems is when you need to either a) CYA or b) reduce headcount

Bite selection is akin to bin packing, now that I think about it.

You missed the killer feature, even more diffusion of responsibility.

"internet chum" is a good one, it echoes "slop bowl".

Essential is contextually defined by whoever implemented the that part of the front-end basically.

It not a killer feature, granted. I would be willing to bet that the cost of the engineering to develop and support this feature as a default capability for the fleet of all vehicles would be less than the value of energy saved ammortized over the lifetime of all relevant vehicles.

Intuitively appreciating that these "boring fundamentals" are the default bottlenecks is a aign of senior+ swe capability.

If it wasn't sold as the next big buzzword you could easily hire 10 new SDR employees who would sell it that way.

For the type of buyer you describe this vehicle parked in the garage, to speculate, may be capable of doing double duty as an automated battery backup for the estate nearby to store energy during times of excess grid capacity and to discharge during periods of high demand or grid interuptions. I would be interested to know if the vehicle includes this capability, or if it could be easily modified to offer this capability. Probably is preferable to an onsite diesel generator for example even if it is not an exactly comparable situation, just due to lower local emissions.

To clarify, even if it is not strictly "spying" by some particular definition, the scope and scale is so large, and the channels to direct actual "spying" resources towards potentially relevant targets that are unveiled through OSINT methods really blur the lines.

Not sure I would agree with the second statement, OSINT is spying, different mechanism but same intent.

I would like to have the opportunity to consider a decentralized consensus algorithm that could accommodate nation state adversaries regularly. Not simply something cryptographically secure and distributed but something which can retroactively route around nodes who are temporarily bad due to external circumstances.

Thank you for validating my decision to check the comments first.

As someone on a team with a less stringent code review culture, AI generated code creates more work when used indiscriminately. Good enough to get approved but full of non-obvious errors that cause expensive rework which only gets prioritized once the shortcomings become painfully obvious (usually) months after the original work was “completed” and once the original author has forgotten the details, or worse, left the team entirely. Not to say AI generated code is not occasionally valuable, just not for anything that is intended to be correct and maintainable indefinitely by other developers. The real challenge is people using AI generated code as a mechanism to avoid fully understanding the problem that needs to be solved.

Slightly harder… if you are naive

I’m certain it’s a memory leak, just not sure solving the problem is worth my time.

My m1 16gb pro gets throttled (and manually restarted) every time vs code hits 100gb ram usage when running jest unit tests. Don’t know who to blame but 99% sure the problem is my source code and not the machine.

Too little too late