LLM Orchestration — Constrained System

Only for registered members United States

7 hours ago

Default job background
$5,000 - $0 (USD) budget
The problem · A quant is awake at 2am. · He types: · "Give me 1s and 0s for when the Big 7 tech stocks are in regime 1 but the rest of the NDX 100 is in regime 0." · When are the generals marching but the army isn't following? · A strategist is reading on a Sunday morning. · She ...
Job description
The problem
A quant is awake at 2am.

He types:
"Give me 1s and 0s for when the Big 7 tech stocks are in regime 1 but the rest of the NDX 100 is in regime 0."
When are the generals marching but the army isn't following?

A strategist is reading on a Sunday morning.

She types:
"Combine Gold, TLT, and VIX into a single fear index. Give me 1s and 0s."
One number that tells her when all the fear assets agree.

A desk doesn't trust short timeframes.

They type:
"BTC on 5-minute bars says 1. BTC on daily bars says 0. Flag every moment they disagree."
When is the short-term lying about the long-term?

Someone gets noisy output.

They type:
"Too noisy. Run it again until stable."
Recursive self-application until the signal stops changing.

This is not a speculative build. The underlying system is complete and deployed.

What exists
— A Python library of deterministic functions. Frozen. Verified. Done.
— The functions share a uniform interface. They require no parameters. They encapsulate their own alignment, warmup, and resampling. You call them, you get a result. Same input, same output, every time.
— The math is not your problem. The math is finished.

What you're building
— An orchestration layer between natural language and this library.
— Parse intent. Route to functions. Compose outputs. Evaluate stability. Return results.
— Not a chatbot. Query in, structured result out.

Technical shape

— Python library of deterministic functions (complete)
— LLM layer:
intent parsing → structured function invocation
— No fine-tuning. No learning loop. No memory.
— Stateless execution — each query independent
— Outputs must remain invariant


Core challenges:
— Mapping ambiguous language to exact function calls
— Composing multiple function outputs into a single result
— Detecting and enforcing output stability (convergence)
— Rejecting any intent that cannot map to the frozen library


Definition of done:
— Natural language queries reliably map to valid function compositions
— Invalid or ambiguous queries are explicitly rejected
— Repeated execution of the same query yields identical results
— Multi-step compositions converge to a stable output when required

What makes this different
— Most LLM orchestration rewards flexibility. This one rewards restraint.
— The system works because of what it doesn't do. The functions don't learn. The interface doesn't change. The architecture doesn't grow.
— Frozen means verified. Verified means attempts to improve it break it.

The real risk
Most systems fail by accumulation. This one fails the same way.

Every addition breaks it:
— Memory? Breaks statelessness.
— Learning? Breaks verification.
— Caching? Introduces hidden state.
— Flexibility? Breaks the constraint.
The architecture survives because someone removes what engineers want to add.
I'm looking for someone who knows when to say no.

The filter
No portfolio. No past work. Answer these three questions.


Question 1:
You're building an orchestration layer for a library of deterministic functions. A senior engineer proposes adding a caching layer to speed up repeated queries.
What is your first question before responding?


Question 2:
A user requests a feature that would require the underlying functions to "learn" from feedback.
The functions cannot be changed.
How do you respond?


Question 3:
You've been on the project three months. You realize a simplification that would eliminate an entire component — but it means admitting the component was unnecessary.
What do you do?

What I'm reading for

Question 1:
Whether you ask "why" before "how."

Question 2:
Whether you protect constraints or engineer around them.

Question 3:
Whether you protect the system or protect your ego.

Compensation
Project-based.
Senior-level budget.
If there's strong alignment, this extends into a longer-term role.

Contract duration of 1 to 3 months.

Mandatory skills:
Large Language Model, Prompt Engineering, AI Model Integration, API Integration, ChatGPT API Integration, Python


Similar jobs

  • Work in company

    Senior Embedded Software Engineer

    Only for registered members

    Our client, a leader and growing presence in the robotics and autonomous vehicles space, is seeking a senior Embedded Software Engineer to deploy and optimize AI capabilities on constrained, real-time hardware platforms supporting autonomous robotic systems. · ...

    Salt Lake City

    1 month ago

  • Work in company

    Senior Embedded Software Engineer

    Only for registered members

    Our client is seeking a senior embedded software engineer to deploy and optimize AI capabilities on real-time hardware platforms. · ...

    Mendon

    1 month ago

  • Work in company

    Firmware Engineer

    Only for registered members

    We are building a series of new devices that give consumers exciting new abilities. · Work at the intersection of ux and hardware to bring our devices to life with low-level code in power and latency-constrained environments. · Integrate wireless connectivity (Bluetooth, BLE, Wi- ...

    San Francisco

    1 week ago

  • Work in company

    SOC Verification Engineers

    Only for registered members

    Responsible for ASIC design verification for various processing blocks within a SOC. · Develop and complete test plans for cache coherency verification of ASIC-based SoCs using UVM-based environments. · Design and implement constrained-random and directed System Verilog testbench ...

    San Jose

    1 week ago

  • Work in company Remote job

    Embedded Systems engineer

    Only for registered members

    We are seeking an Embedded Systems Designer to develop and integrate compact electronic systems for innovative hardware products, with a focus on circuit design, PCBA development, and flexible PCB technologies. This role involves designing schematics and multilayer boards, workin ...

    $3,500 - $0 (USD) budget

    1 day ago

  • Work in company

    Server Power and Performance Attainment Lead

    Only for registered members

    We are seeking an experienced Server Power Performance Attainment Lead with expertise in overall SoC power constrained performance. · Advance state of methodology for power constrained performance achievement using power models. · Work with business unit to define product configu ...

    Austin Full time

    1 month ago

  • Work in company

    Embedded Software Engineer IV – AI Applications

    Only for registered members

    Our client is seeking a senior Embedded Software Engineer to deploy and optimize AI capabilities on constrained hardware platforms supporting autonomous robotic systems. · ...

    Lehi

    1 month ago

  • Work in company

    Embedded Software Engineer IV – AI Applications

    Only for registered members

    Our client a leader and growing presence in the robotics and autonomous vehicles space is seeking a senior Embedded Software Engineer to deploy and optimize AI capabilities on constrained real-time hardware platforms supporting autonomous robotic systems. · DeployandoptimizeAI/ML ...

    Mendon

    1 month ago

  • Work in company

    Robotics Engineer

    Only for registered members

    Design and deploy modular sensor stacks across multiple machine types. Build low-latency AI pipelines under constrained bandwidth and compute environments. · ...

    San Francisco

    2 weeks ago

  • Work in company Remote job

    Senior Full-Stack Developer

    Only for registered members

    We're building a business operating system where AI is a core system component. We're looking for a senior developer who knows how to put guardrails on AI: structured inputs constrained outputs explainable decisions human approval loops safe defaults This is a full-time freelance ...

    $15 - $25 (USD) per hour Full time

    2 weeks ago

  • Work in company

    Software Architect

    Only for registered members

    Algorized is a VC-funded Silicon Valley deep-tech company building edge-AI models that give robots real-time human awareness using existing wireless sensors. · Lead the design and development of software architecture across embedded, backend, and frontend systems. · Create and ma ...

    Campbell

    2 weeks ago

  • Work in company

    Embedded Software Engineer IV

    Only for registered members

    Dice is the leading career destination for tech experts at every stage of their careers. Our client, ConsultNet, LLC, is seeking senior Embedded Software Engineer to deploy and optimize AI capabilities on constrained real-time hardware platforms supporting autonomous robotic syst ...

    Lehi

    1 month ago

  • Work in company

    Embedded Software Engineer IV

    Only for registered members

    Dice is the leading career destination for tech experts at every stage of their careers. Our client, ConsultNet LLC seeks an Embedded Software Engineer IV to deploy and optimize AI capabilities on constrained real-time hardware platforms supporting autonomous robotic systems. · ...

    Mendon

    1 month ago

  • Work in company

    Senior Electrical Engineer

    Only for registered members

    We are a fast-growing defense startup developing advanced autonomous turret systems that combine robotics, computer vision, and precision control to detect and neutralize small drones and loitering munitions. · Design and develop electrical systems, including schematics and PCB l ...

    Austin

    3 weeks ago

  • Work in company

    Server Power and Performance Attainment Lead

    Only for registered members

    We are seeking an experienced datacenter/Server Power and Performance Attainment Lead with expertise or significant interest in overall SoC and IP level power constrained performance. · ...

    Austin

    3 weeks ago

  • Work in company

    Server Power and Performance Attainment Lead

    Only for registered members

    We are seeking an experienced datacenter/Server Power and Performance Attainment Lead with expertise or significant interest in overall SoC and IP level power constrained performance. · Undergrad degree required. Bachelor's or Master's degree in related discipline preferred · ...

    Austin Full time

    3 weeks ago

  • Work in company Remote job

    System Engineer – Atmospheric Water Generator

    Only for registered members

    This role involves developing a deployable Atmospheric Water Generation (AWG) platform addressing water scarcity under climate-constrained environments. The candidate will operate at the level of system physics definition, not software operation. · Constructing and validating the ...

    4 weeks ago

  • Work in company

    Software Architect

    Only for registered members

    Position in firmware design with strong proficiency in C/C++ for embedded systems and expertise in memory management. · ...

    Minneapolis

    1 week ago

  • Work in company

    GPU Design Verification Engineer

    Only for registered members

    Help develop comprehensive verification plans, define verification goals, coverage metrics, and sign-off criteria aligned with architectural specifications and integration requirements. · Architect and implement advanced verification environments by designing, developing, and mai ...

    Austin

    2 weeks ago

  • Work in company

    Senior Firmware Engineer

    Only for registered members

    Great opportunity to join a well-established startup expanding into new markets. · ...

    San Francisco

    1 month ago