Product & Software Engineering
Frontend, backend, DICOM, FHIR, workstation integrations. IEC 62304 Class C where it has to be.
End-to-end SaMD delivery
AI SaMD from concept to clearance in as little as 9 months.
Most teams stall at the hand-off between R&D, quality, and regulatory. We collapse those gaps into a single delivery, so your device ships while your competitors are still scoping vendors.
We Did It All: R&D, Clinical Study, and 510(k) Cleared in 12 Months
Built and cleared an AI cardiomegaly detector end-to-end: algorithm R&D to 510(k) clearance in under 12 months.
We went from idea to FDA clearance in 18 months and first Install for AI Tumor Tracking Software
Prototype → 510(k) → QMS → FDA audit passed. Multi-year partnership with ongoing development.
Dental CADe Concept to 510(k). We built the app, trained the AI, ran the MRMC study, cleared the FDA.
We built the app, trained the AI, ran the MRMC study, and cleared FDA.
Competitor: 5 years, $500k, no results. Us: 18 months, FDA cleared, IPO filed.
Full-cycle development of GyriCalc, a first-of-its-kind pediatric neuroimaging diagnostic. FDA cleared (K250686) July 2025. Filed for IPO in June 2025.
We helped build the CBCT module, got it FDA cleared, and supported them through the Varian acquisition.
Multi-Year Partnership: CBCT Development → FDA Clearance → Maintenance → Support → M&A Support
We Architected Neosoma's Fast Lane to FDA: A Modular Strategy Built for Rapid Repeat Clearances
We isolated Neosoma's new CNN for FDA review, won clearance in 3 months, and built a framework for every product after.
We Helped Build OCT Software and Get It FDA Cleared
UI design, Custom C++, Qt, Python development, IEC 62304 and FDA software documentation for a successful 510(k) submission
We Turned a MATLAB Prototype Into 510(k)-Ready Software With Full Documentation
MATLAB to Python port with performance improvements, documentation, and verification
We resolved a hold letter that threatened a clinical study redo and got it 510(k) cleared on Christmas Day.
Resolved FDA hold letter that would have required repeating the entire clinical study. Cleared shortly after.
Recent end-to-end engagements
What 'one roof' actually means
You hand us the data. We hand you a clearance letter.
Most programs fail in the seams. Engineering ships something the regulatory team can’t defend, or regulatory scopes something engineering can’t build. We don’t have seams. Software engineers, AI/ML scientists, MDs, PhDs, and regulatory leads on the same team, on the same Slack, on the same deadline.
You give us the clinical intent, the data, and the target FDA center. We give you a cleared device. Submission package, V&V evidence, cybersecurity dossier, labeling, the whole design history file. 45+ submissions and counting.
What 'end-to-end' actually includes
One accountable team. Eight workstreams. One cleared device.
Wrangling multiple vendors is like herding cats. But unlike cats, your company doesn’t have nine lives. You’ve got one shot with FDA. Make it count.
AI/ML Model Development
Data curation, training, performance, generalization, MRMC reader studies. PyTorch, TensorFlow, MONAI.
Quality Management System
ISO 13485 QMS that holds up under an FDA audit. Stand up a new one, or plug into yours.
Regulatory Strategy & Submissions
510(k), De Novo, Pre-Sub, PMA supplements, Breakthrough Device Designation. We write, we submit, we respond.
Verification & Validation
Unit, integration, system, clinical performance. Traceable requirements to test evidence.
Cybersecurity (FDA 2023)
SBOM, threat modeling, penetration test coordination, SPDF, vulnerability handling. The new guidance, done right the first time.
Risk Management (ISO 14971)
Hazard analysis, FMEA, benefit-risk. Tied into design, V&V, post-market from day one.
Post-market & PCCP
Predetermined Change Control Plan, post-market surveillance, MDR reporting, SaMD model monitoring.
Not in scope: running your clinical trials, your wet lab, or your sales force. Everything else between your data and a clearance letter is ours.
Proof, not promises
Nine engagements. Engineering and regulatory under one roof.
Every case study below had Innolitics on both the engineering side and the regulatory side. No vendor handoffs, no blame boundary. One team, one outcome.
Voice of the customer
What founders say after we ship.
The lines CEOs repeat on reference calls.
I needed a software development partner to write the software, train the AI, and get FDA clearance. An investor once told me that it would take me $5 million and 5 years to get to where we are now. Innolitics got me here 3 years ahead of schedule and $4 million dollars under budget.
Innolitics are more than a group of top SaMD engineers and reg consultants—they are trusted advisors and are like family. They are a one stop shop for AI/ML algorithm R&D, full stack web development, and FDA regulatory clearance. We built SmileDx, a dental CADe from scratch all without needing to raise external funding and within a reasonable timeframe.
After spending $500K over five years on consultants who knew how to sell but couldn't execute, Innolitics took us from concept to FDA clearance in under 18 months for a first-of-its-kind MRI neuroimaging device. When FDA questioned our original approach, they adapted without losing momentum. We submitted on schedule and cleared in 4.5 months. I've seen many engineers at J&J and Ethicon — Innolitics is many standard deviations ahead of the curve.
Innolitics played a vital role in the development of our software according to our quality management system, as well as the establishment of our engineering team. Their contributions were integral to fostering a culture of excellence for our complex, software-only medical device. We have had the pleasure of collaborating with them since 2018 and wholeheartedly advocate their services.
Our first FDA clearance was monolithic — every component submitted as one device. When we needed to clear a second AI indication, we assumed we'd have to re-submit the entire platform. Innolitics saw something we didn't: a modular regulatory architecture that isolated the new CNN as the only component under review. That single insight saved us months of redundant documentation and positioned every future tumor type as a lightweight module swap. We submitted in September 2025 and cleared in December — 3-month review, zero major deficiencies. I've worked with regulatory consultants before. Innolitics is the first one that understood our software well enough to design a regulatory strategy from the architecture itself.
Innolitics is an incredible partner and consistently surpasses our expectations. They have an extremely agile team, adapting to our needs across back-end and front-end tasks seamlessly. When we needed support around ISO 62304 compliance for FDA requirements, they jumped right in and provided us compliant documentation. They also assisted us as we developed a regulatory strategy around FDA Cybersecurity and HIPAA Compliance. The Innolitics team is efficient, fair, and highly ethical. They are an absolute pleasure to work with.
We recently got FDA cleared and Innolitics' responses were absolutely clutch to craft the strategy that finally worked. Unlike other consultants, who wanted us to do more work and spend more money on clinical testing, Innolitics found a path of least resistance using a combination of our existing validation and thoughtful responses to FDA. We received the best Christmas present ever – our 510(k) clearance letter.
Dr. Andrew Smith, MD PhD
Co-Founder at AI Metrics &
Chair of Diagnostic Imaging at St. Jude Children's Research Hospital
Your product is probably in here
If it has an algorithm, we've probably already done it.
Six SaMD archetypes below. Radiology AI, imaging, signal-processing, surgical, IVD, remote monitoring. Pick the row that looks most like yours.
Radiology AI
Radiology CADe, CADx, and segmentation
Unlocks: Detection, quantification, or measurement software against imaging. Reader studies, MRMC, AI/ML validation, 510(k) with algorithm-specific SOUP controls.
Imaging
Pathology, ophthalmology, and multi-modal imaging
Unlocks: Whole-slide image viewers, OCT analyzers, fundus AI. Color calibration, DICOM, clinical workflow integration.
Signal
Signal-processing SaMD
Unlocks: ECG, EEG, PPG, cardio monitors. DSP algorithms, foundation models on physiological signals, 510(k) and De Novo pathways.
Surgical
Surgical planning and navigation
Unlocks: Pre-operative planning, intra-op guidance, 3D reconstruction. Cybersecurity and human factors scrutiny the FDA actually reads.
IVD
Point-of-care IVD software
Unlocks: Assay result interpretation, LDT-to-IVD migrations, CLIA-waived device software. CDRH and CDER crosswalks when the asset is drug-linked.
RPM
Remote patient monitoring and wearables
Unlocks: Device-to-cloud pipelines, mobile apps, clinical dashboards. Part-11, HIPAA, and cybersecurity built in, not bolted on.
How we engage
Turnkey, or rescue. You pick.
Same team, same quality system, same FDA reviewers. We take the wheel, or we take it back.
Option A
Turnkey end-to-end
We own the outcome from concept to clearance letter. Fixed scope, fixed timeline, fixed price where we can. If we miss the date, we make it right. This is the ‘we handle everything’ option.
Best for: Founders and CEOs who want a cleared product, not a project plan. Teams without internal SaMD engineering or regulatory depth.
Option B
Rescue and accelerate
You have a team, a codebase, maybe a failed submission or a hold letter. We come in, triage, and take engineering + regulatory to the finish line. No finger-pointing, no restart. We meet the code and the submission where they are.
Best for: Programs that stalled after a negative FDA interaction, a consultant handoff that didn't hold, or an internal team that ran out of runway before the finish line.
Recognize your product above?
Tell us what you're building. We'll tell you what it takes.
Straight talk in 30 minutes: scope, timeline, team, references, what we would and wouldn’t do.
Before you book the call
What founders and CEOs ask on the first call.
What does 'end-to-end' actually mean? What's in scope?
Everything between your clinical intent and a clearance letter. Product and software engineering, AI/ML model development, quality system, regulatory strategy and submission, V&V, cybersecurity, human factors, risk management, and post-market. What’s not in scope: your clinical trials, your wet lab, your sales force, and the things you’d rather keep in-house. We map the line on the first call.
Do we have to bring our own algorithm, or can you build the model too?
Either. AI Metrics and SmileDx both started without a working model; we trained them. PhotoniCare had a working optical system and needed the software; we wrote the software. Neosomas had a cleared first device; we redesigned their regulatory architecture so every future AI indication is a module swap. We meet you where you are.
How do you price, and what happens if you miss the timeline?
Fixed-fee milestone pricing where we can scope cleanly; time-and-materials where scope is genuinely unknown. We tell you the tradeoff before you sign. If we miss a committed date on a fixed-scope milestone, we make it right. We put that in writing.
Who owns the code, the model weights, and the submission artifacts?
You do. Day one. Source code in your repos, model weights in your infrastructure, design history file in your QMS, submission under your regulatory lead’s signature. We’re the team that builds it. You’re the sponsor.
We already have engineers. Why not just hire more?
Because the bottleneck isn’t engineering capacity, it’s the integration between engineering, regulatory, AI/ML, and cybersecurity. Hiring fixes one of those. We bring all four as one team with shared incentives. If you want, we help you hire a real in-house team on the way out, the way we did for AI Metrics.
Our last consultant handed us a binder. What do we actually get?
A cleared device. Working code in your repos, model artifacts you can retrain, a QMS that passes an FDA inspection, a design history file a reviewer can navigate, V&V protocols and reports, the submission package, the cybersecurity dossier, and a post-market plan with a Predetermined Change Control Plan already drafted. Not a binder. A product.
Can you help us fix a failed submission or an FDA hold letter?
Yes. SimBioSys came to us with a hold letter; we found the response that landed the clearance without a new clinical study. Neosomas came to us with a monolithic first submission and wanted a better architecture; we designed one. Rescue engagements are Option B in the engagement models above.
How long does an end-to-end engagement take?
The AI Metrics number was 18 months, idea to clearance. GyriCalc was a similar magnitude. If you are well-funded and already have the training data in hand, we can compress the whole thing into 9 months. Rescue engagements can close faster still because most of the work is already done. We give you a real number on the first call, anchored against a specific FDA center and pathway, not a generic range.
Depth if you want it
What we write about end-to-end SaMD.
Innolitics editorial on AI/ML validation, FDA pre-submissions, and the engineering velocity that makes the timeline possible.
Definitive Guide to AI/ML SaMD Acceptance Criteria
Setting acceptance criteria incorrectly (too high or too low) can delay your FDA submission by weeks or trigger rejection. This guide analyzed 784 ...
Read more →
FDA Pre-Subs: Best Practices, FAQs, and Examples
A Pre-Sub is a mechanism for requesting formal written feedback from the FDA, and (optionally) a one-hour meeting. Pre-subs are a useful means to m...
Read more →
Frustrated with Slow AI Adoption? Here’s why.
AI-assisted medical device development is FDA-compliant when done right. The obstacle isn't regulation. It's the messy codebase you've been ignoring.
Read more →Ready to move
30 minutes to a go / no-go.
You’ll talk to someone who has cleared this class of device before, in this FDA center, on a deadline that mattered.
Let's Talk
Every great partnership starts with a conversation. Fill out the form below for a discovery call, and an Innolitics team member will contact you soon.
I learned more from an hour-long Innolitics sales call than an entire paid engagement from other consultants.
Richard Clark
Executive in Residence at Washington University