Case studies

Execution intelligence, in real sessions.

Problem → Prediction → Outcome — from measured LOCI sessions, not testimonials.

Embedded · BLE firmware

A dead path in a clean build

Caught pre-merge
  1. Problem

    A BLE sensor gateway was committed "feature complete". The build was clean, but the poll-tick and reconnect events fell through to default: break — they never ran.

  2. Prediction

    LOCI's preflight flagged the plan with a safety fail before the edit landed, from the firmware binary — no board, no probe, no power analyzer.

  3. Outcome

    The dispatch fix passed LOCI's post-edit check before merge. The timeline was modelled; nothing was flashed.

Cloud · OpenSSL token service

A p99 regression caught at PR

18× p99, prevented
  1. Problem

    A change encrypted session tokens with AES-256-GCM by creating a new EVP cipher context on every request.

  2. Prediction

    LOCI predicted p99 latency 18× over budget — per-request alloc/free on the hot path, throughput regressing — before a single test ran.

  3. Outcome

    The agent pooled the context and reused it per call. The throughput regression never reached production.

AI · CUDA paged-attention

A GPU spill, predicted from the binary

1.32× faster, within budget
  1. Problem

    An agent raised GEMM register tiling on SM_90 to reuse more data — a change that looks faster in source.

  2. Prediction

    LOCI predicted register pressure over budget (132 > 128 regs) — ptxas would spill — read from PTX / SASS, before launch.

  3. Outcome

    Re-tiled to single-buffered K. Within budget, 1.32× faster — no spill, no slow inference.

Robotics · real-time motor control

A control-loop deadline, caught before deploy

1 kHz loop, deadline held
  1. Problem

    An agent added a smoothing filter to a robot arm's motor-control loop to reduce servo jitter — a change that looks harmless in source.

  2. Prediction

    LOCI predicted the loop's worst-case execution would exceed the 1 ms control period on the target MCU — the deadline miss and jitter that would destabilize the servo — read from the binary, before it ran on hardware.

  3. Outcome

    The agent moved the filter off the hot path and precomputed its coefficients. The 1 kHz loop stayed within deadline — stability preserved before the robot ever moved.

Automotive · autonomous navigation

Route to the point, ready to go

planner on the control cycle
  1. Problem

    An agent wrote the route planner that turns a destination into waypoints for the navigation stack — refreshing the map as the vehicle drives.

  2. Prediction

    LOCI predicted the planner's worst-case would overrun the ECU's update cycle — waypoints reaching the map late and stale — read from the binary, before a test drive.

  3. Outcome

    The agent bounded the search and cached the map tiles. The route now computes inside the control cycle — the right data reaches the map on time, ready to drive.

Real measured LOCI sessions · predictions verifiable on real silicon.