Fixed-price packages · Booking now
Four fixed-price engagements with a clear scope and a known price.
Some problems are common enough that we can scope them before the first call. These packages cover four of them. Each one has a set timeline, a written deliverable you keep, and a price you know before we start.
At a glance
The packages
Rust Migration Assessment
A two-week audit of a C or C++ codebase with a risk map, memory-safety findings, and a phased plan for moving to Rust.
Starting at $14,500 · 2 weeks
Cloud & GPU Cost Audit
Find the waste in your cloud bill, Kubernetes clusters, and inference serving, with a savings report and a roadmap to act on it.
Starting at $6,500 · about 2 weeks
Hardware Test Automation
Build a new test rig or modernize an old one for boards, devices, and firmware, in phases you approve one at a time.
Starting at $6,500 (design phase)
AI-Built Codebase Rescue
A one-week review of an app built quickly with AI tools, covering security, reliability, and scaling, with a ranked fix plan.
Starting at $4,500 · 1 week
Final scope and price are set on a free scoping call.
Package 01
Rust Migration Assessment
Two weeks with your C or C++ code. You get a clear picture of where memory-safety risk lives and a migration plan you can schedule one phase at a time.
Who it is for
Engineering leads in defense, medical devices, automotive, and aerospace who maintain C or C++ and are hearing questions about memory safety from customers, regulators, or their own security team. It also fits teams deciding whether new components should be written in Rust next to existing code.
The problem
Memory-safety bugs are behind many of the most serious vulnerabilities in systems software, and more customers and agencies now ask suppliers for a plan to address them. A full rewrite is rarely realistic. You need to know which parts of the code carry the most risk, what Rust would take in each place, and which order makes sense.
What is included
- Kickoff and architecture walkthrough with your team
- Static analysis and sanitizer runs where your build allows (clang-tidy, cppcheck, AddressSanitizer, UBSan)
- Manual review of the highest-risk code: parsers, network and file input, IPC, buffer and pointer handling
- Toolchain and target review: compilers, RTOS or BSP, cross-compilation, and any certification standard you work under (DO-178C, IEC 62304, ISO 26262)
- A build spike to confirm Rust compiles and links for your target, with a proposed FFI boundary (bindgen or cxx)
Deliverables
- Risk map. Every module rated for exposure to untrusted input, defect history, change rate, and migration difficulty.
- Memory-safety findings list. Each finding with location, CWE class, severity, and a recommended fix in the current language or in Rust.
- Phased migration plan. Phase order, interface boundaries, test strategy, effort ranges, and staffing.
- Readout session for engineering and leadership.
| Timeline | 2 weeks from code access to readout |
|---|---|
| Price | Starting at $14,500. Final scope and price are set on a free scoping call. |
| What we need from you | Read-only repository access, or a review on your own hardware if the code cannot leave your site. Build instructions or a working build environment. A technical contact for about two hours a week. Any existing bug, CVE, or static-analysis history. An NDA if you need one. If the code is export-controlled, tell us at the start so we can agree on how and where the review happens. |
Package 02
Cloud & GPU Cost Audit
We read your bill, your clusters, and your inference stack, then tell you what you are overpaying for, what each fix is worth, and what order to do them in.
Who it is for
Companies running on AWS or Azure, usually with Kubernetes in production, whose spend has grown faster than their usage. Especially teams serving machine-learning or LLM inference on GPUs, where idle or oversized capacity gets expensive quickly.
The problem
Workloads request far more CPU and memory than they use. GPUs sit half idle between batches. Old volumes, snapshots, and test environments keep running. Commitments were sized once and never revisited. Nobody can say which team or feature drives the bill.
What is included
- Spend breakdown by service, account, environment, and team
- Kubernetes review: requests and limits against real usage, autoscaling, node pools, and spot or low-priority capacity
- Inference serving review: GPU utilization, batching, model and instance sizing, autoscaling, cold starts, and cost per request or per million tokens
- Idle and orphaned resources, storage tiers, data egress, and log and metric volume
- Savings Plans, Reserved Instances, and Azure reservations sized to your steady load
- Guardrails: tagging, budgets, and anomaly alerts so savings stay saved
Deliverables
- Savings report. Every finding with estimated monthly saving, effort, and risk.
- Quick-win list of low-risk changes your team can make right away.
- 30/60/90-day roadmap for the larger changes.
- Walkthrough with engineering and finance.
| Timeline | About 2 weeks. Roughly three hours of your team's time in total. |
|---|---|
| Price | Starting at $6,500. Final scope and price are set on a free scoping call. |
| What we need from you | Read-only billing access (AWS Cost and Usage Report or Azure Cost Management export), a read-only cloud role, read-only cluster access, your infrastructure-as-code repositories if you have them, and one contact for a kickoff and a walkthrough. Access can be revoked the day the audit ends. |
Package 03
Hardware Test Automation
A test rig and test software for your boards, devices, and firmware, built in three phases. Each phase is scoped before it starts, and you decide whether to continue after each one.
Who it is for
Hardware and embedded teams that still test by hand, manufacturers running an aging test station that one person knows how to keep alive, and product teams getting ready for their first production run.
The problem
Manual tests are slow and vary from one operator to the next. Results end up in spreadsheets and screenshots. Older stations run on unsupported computers and software, and nobody wants to touch them. When volume goes up, test becomes the bottleneck.
Test plan & rig design
1 to 2 weeks
- Review of the device, the current test process, and the instruments you own
- Test plan with requirements, limits, and coverage
- Rig architecture: instruments, switching, fixture, and interface board if needed
- Bill of materials with costs
- A fixed scope and price for Phase 2
Pilot station
4 to 6 weeks
- One working station: fixture, wiring, or interface PCB
- Instrument and device control over SCPI, serial, CAN, GPIO, and JTAG/SWD for firmware loading
- Test runner in Python, or Rust where speed or reliability calls for it
- Pass/fail limits, per-unit records, and a simple operator screen
- Validated on known-good and known-bad units and compared against your manual results
Production & scale
Scoped after the pilot
- Additional stations
- Database or MES integration
- Hardware-in-the-loop tests in your firmware CI pipeline
- Operator training and ongoing support
| Deliverables | Test plan, schematics and bill of materials, all source code, an operator runbook, sample test reports, and a handover session. You own everything we build. |
|---|---|
| Timeline | Phase 1 in 1 to 2 weeks. Phase 2 in 4 to 6 weeks, depending on parts lead times. |
| Price | Starting at $6,500 for the design phase. Final scope and price are set on a free scoping call, and each later phase is priced before it begins. |
| What we need from you | Sample units, including known-bad ones if you have them. Schematics, test points, and any existing test procedures. A list of instruments you already own. Bench access, or the ability to ship units to us. One engineering contact. |
Our team has built factory-floor test automation for prototype memory devices in semiconductor manufacturing, and our engineering lead holds a degree in Electrical Engineering.
Package 04
AI-Built Codebase Rescue
Your app was built quickly with AI coding tools and now it has real users, real data, or a launch coming. We spend a week finding what will break and hand you a plan to fix it.
Who it is for
Startups and internal teams whose product was built largely with tools like Cursor, Copilot, Claude Code, Lovable, Bolt, or Replit. It is a good fit before a launch, a funding round, a customer security review, or when production has started to wobble.
The problem
The app works in a demo, but there are things nobody has checked: keys committed to the repository, authorization that only exists in the browser, no tests, deploys that break, and queries that will slow to a crawl at ten times the users. It is hard to tell which of these matter most.
What the review covers
- Security: secrets in code and git history, authentication and authorization on every route, input validation, dependency vulnerabilities, OWASP Top 10
- Reliability: error handling, data integrity, migrations, backups, tests, CI/CD, logging and monitoring
- Scaling: database queries and indexes, caching, background jobs, hosting configuration, and running costs
- A repair-or-rebuild recommendation for each major area
Deliverables
- Findings report ranked by severity, with file and line references.
- Prioritized fix plan with an effort estimate for each item. You can hand it to any developer.
- Walkthrough call with your team.
- A fixed scope and price for the optional implementation sprint.
| Timeline | Review in 1 week. Optional sprint in 2 weeks. |
|---|---|
| Price | Starting at $4,500 for the review. Final scope and price are set on a free scoping call. |
| The optional sprint | After the review, we can fix critical and high findings in priority order: rotate exposed secrets, enforce authorization on the server, add CI with tests on your critical paths, set up error tracking and uptime monitoring, and leave a written handover of everything we changed. |
| What we need from you | Read access to the repository (pull-request access for the sprint), details of where and how it is hosted, a short description of what the app is supposed to do, and about an hour for kickoff and the walkthrough. |
How packages work
Simple terms
Price first
Final scope and price are set on a free scoping call. You get a short written scope with the fixed price before any work begins. If something outside the scope comes up, we tell you and quote it separately.
Senior-led
Every package is led by a senior engineer, with engineers on our team handling parts of the analysis and build. A senior engineer reviews every deliverable before it reaches you.
Our experience
Our team brings six years of production Rust, plus Go, AWS, Azure, and Kubernetes. We have run release and platform engineering for a large enterprise, shipping up to 50 releases a week without a failed release in three years. We have built factory test automation in semiconductor manufacturing, and our engineering lead holds a degree in Electrical Engineering.
You keep the work
Reports, plans, source code, and design files are yours. If you decide to carry out the plan with your own team or someone else, that is a perfectly good outcome.
Start with a short email
Tell us which package you are considering and a little about your system. We will reply within one business day with a straight answer about fit and price.