ROLE-001 · ENGINEERING · 2026
High Frequency Software Engineer (HFSWE)
Build the infrastructure, orchestration, and internal platforms for a high-frequency software company optimized for speed, judgment, and process engineering.
The role
We think software is entering a high-frequency phase. Code is abundant. The scarce things are judgment, latency, process design, and the infrastructure that lets a small team run many bets at once without collapsing into chaos. This role is about building the machine that builds the machine: signal pipelines, orchestration layers, internal tooling, replay systems, and the runtime guarantees that let us move at comical speed while still producing work that can be trusted. We want engineers who can operate at software-abundance scale: not just shipping more code, but steering many concurrent workstreams, agents, and experiments with unusually high leverage. More importantly, we think the middle of software work is disappearing. The hard part is shifting from typing code to shaping intent, coordinating agent execution, and building systems that can adapt to real users rather than averages. A lot of this role is judgment engineering: turning experience, constraints, and scar tissue into systems that can move fast without becoming reckless.
01
What You'll Build
- Signal-ingestion and orchestration systems that turn incoming information into experiments, workflows, decisions, and product changes quickly.
- Model-agnostic runtime infrastructure that routes the best model to each task while keeping the overall system verifiable and reliable.
- Internal platform primitives for high-frequency software: deployment, observability, isolation, feedback loops, and experiment management.
- Replay, backtesting, and evaluation systems that tell us whether an AI-assisted workflow would actually have improved outcomes.
- Systems that help software adapt to actual user behavior and real workflows instead of forcing everyone into the same frozen interface or process.
- Infrastructure that lets the company run a rapidly evolving portfolio of bets rather than one narrow software roadmap at a time.
02
How We Think
- The moat is not generic AI access. The moat is process engineering: understanding how real work gets done and encoding that understanding into durable, load-bearing systems.
- We optimize for comically low latency, but not for speed at the expense of correctness. Reliability is part of the product.
- We have high conviction about important problems and low conviction about the exact solution. The system should let us test and adapt fast.
- We care more about building the machine that builds and improves systems than about heroics on individual commits.
- We are model-agnostic by design. No one provider gets to become a single point of failure in the stack.
- The center of software work is moving. Forming the right intent, structuring context, and validating outcomes matters more than manually translating every idea into code.
- We optimize for impact, not activity. More output only matters if it improves the system, the product, or the business.
03
How You Work
- You operate like an AI-native engineer: multiple threads, multiple sessions, multiple agents, tight review loops, and high-volume output without losing taste.
- You use CLI coding agents and modern IDE copilots fluently. Codex, Claude Code, Cursor, Antigravity, and adjacent tools are part of your daily workflow, not a novelty.
- You can drive full-stack delivery with AI assistance, but you do not confuse generated output with engineering judgment.
- You are comfortable shipping at software-abundance scale, where a single engineer can move the equivalent of a small team, as long as the surrounding constraints, checks, and architecture are right.
- You understand that as agentic execution gets easier, review, testing, release discipline, and context quality become even more important.
04
Requirements
- Strong systems instincts and comfort around performance bottlenecks, concurrency, and event-driven infrastructure.
- Experience with low-latency services, distributed systems, internal platforms, or similarly demanding runtime environments.
- Strong full-stack engineering judgment and the ability to move between product surface, backend systems, runtime behavior, and infrastructure when needed.
- Clear evidence that you have shipped serious software recently with AI tooling in the loop, not just experimented with it.
- Ideally, substantial engineering experience from before the generative AI shift as well, so you know what good looked like before code generation got cheap.
- Comfort operating in ambiguous product and technical environments where the bottleneck is often figuring out what should be automated, not just implementing it.
- Ability to work from user intent, product context, and structured feedback rather than waiting for perfectly specified tickets.
- Respect for failure semantics, measurement, observability, and operational discipline.
- Good taste around automation: knowing what can be delegated to machines, what must be constrained, and what needs a human in the loop.
05
Example Problems
- Reduce the time from incoming external or internal signal to a meaningful prototype, workflow change, or operator action without sacrificing auditability.
- Build the equivalent of internal AWS for high-frequency software: the infrastructure layer that lets many experiments, agents, and product lines coexist without chaos.
- Design a replay or backtesting harness that evaluates whether an AI system, workflow, or product decision would actually have outperformed our previous process.
- Turn a team-specific workflow into a system that is fast, repeatable, trusted, and capable of improving itself through feedback.
- Build the coordination and context layer that lets humans and agents work against the same intended outcomes instead of losing time in translation.
- Ingest the software equivalent of order flow: customer feedback, product telemetry, competitor moves, and operational signals, then make the system responsive enough to act on them quickly.
Apply
Tell us why this role is the one.
Send a short note and your best evidence of the work to contact@generalliquidity.com. Every position begins with a short paid trial period.