What a Low WER Missed in a Medical Transcript
Our synthetic interview scored about 1.4% WER, while a separate FDA recording contained four FAIRS substitutions that the attention rules did not flag
Technically reviewed by Victor Sineglazov, D.Sc.
Our synthetic interview had a word error rate of about 1.4%. In a separate real recording, the recognizer wrote FAIRS instead of FAERS four times, and none of those segments received an attention flag. The clean aggregate score had not told us whether the terminology checks would catch that error.
These observations came from an October 7, 2026 live check of audio ingestion for a professional learning platform. We tested a known-script recording and a public FDA Grand Rounds recording. Only the synthetic recording had a complete reference script, so the 1.4% result belongs exclusively to that material.
What the synthetic score measured
The synthetic interview lasted about 32 minutes and 24 seconds, spoken by Microsoft Zira Desktop. The pipeline split it into three processing pieces and produced 264 transcript segments, then grouped the text into 21 retrieval fragments. A segment's transcription boundary and a fragment's search boundary were different objects. We measured word error rate against the script after normalizing numbers. The result was approximately 1.4%. The reference mattered: a numeric value written as digits should not automatically count as a different fact from the same value spoken in words. Normalization choices still need to remain available beside the score.
The attention rules flagged 12 segments, or about 4.5% of the synthetic set. Against the scripted reference, recorded precision was 0.58 and recall was 0.23. Much of the apparent error involved number formatting or units, including abbreviated units versus their spoken names.
We did not calibrate the attention thresholds on that text-to-speech sample. Clean synthetic speech could help test ingestion and compare wording with a known script, but it could not represent the full range of uncertainty in real terminology or recording conditions.
A confident terminology error in real speech
The FDA recording lasted 48 minutes and 17 seconds. It produced 356 transcript segments and 30 retrieval fragments. The rules flagged 22 segments, about 6.2%. Those counts describe this one recorded run.
Our spot check found FAERS written correctly three times and as FAIRS four times. None of the affected segments was marked for attention. The same record noted 41 correct occurrences of MoCRA. Correct handling of one specialist term therefore did not tell us how another would behave.
| Material | Recorded observation | What it supports |
|---|---|---|
| Synthetic interview | About 1.4% WER after number normalization. | A score against the known script for that recording. |
| Synthetic attention flags | Precision 0.58; recall 0.23. | The rules' behavior against that limited scripted reference. |
| Real FDA recording | FAERS appeared correctly three times and as FAIRS four times; no affected segment was flagged. | A concrete terminology blind spot in this recording. |
| Real-audio WER | No complete human reference transcript was recorded. | No measured WER for the real recording. |
The attention mechanism was intended to direct inspection toward uncertain output. This failure was outside that useful region: the model's confidence did not expose the misheard acronym. An absence of flags could not be read as an absence of terminology errors.
Transcription, retrieval and citation time can diverge
The synthetic run retrieved the intended moment for ten out of ten questions. Nine answers cited the intended moment; one cited an adjacent segment. On the real recording, all five selected questions retrieved the expected material, while four answers cited the intended moment within the recorded 30-second tolerance.
These were different checks. Retrieval could find a useful fragment despite a misspelled acronym. An answer could then cite a nearby interval that did not contain the relevant spoken fact. A transcript-level score alone would not reveal either outcome. This is why we kept the original recording accessible behind the citation. A reviewer needs to hear the relevant interval, rather than infer transcription fidelity from a well-formed answer.
The presence of a playback control does not establish that someone listened.
A review queue needs evidence of its misses
Our notes prepared a listening list containing five flagged and five unflagged examples, including segments with FAIRS. The developer's listening verdict was not recorded. We therefore cannot present that list as completed human validation or use it to estimate the real recording's error rate.
For a next evaluation, we would first label selected real-audio intervals independently of the attention flag. The sample should include specialist names, abbreviations and measurements alongside ordinary speech. Inspecting only flagged segments would make the FAERS failure hard to discover again.
A terminology list could supply targeted assertions, but it introduces its own ambiguity. A recognizer should not silently replace every instance of a common word with a familiar acronym. Candidate corrections need the recording and context. The observed confusion is evidence for adding a test case, rather than proof of a deployed correction.
We would retain a separate count of critical-term errors and report how often the review queue caught them. Broad WER and triage precision would remain useful, with explicit normalization and sampling. Neither would carry a claim about clinical safety.
Corrections need a different lifecycle from recognition
The live check also exercised a segment edit. The corrected text became human-authored, its attention flag cleared and the retrieval fragment rebuilt. Reindexing preserved that edit without another speech-model call. A separate transcribe-again action warned that it would replace a human correction, then regenerated the transcript.
Those state changes answer a question the score cannot: after a person fixes a term, will an ordinary index rebuild erase the work? In the recorded check, it did not. Explicit retranscription was a different operation.
Our synthetic-corpus article describes the reference-timing checks behind this run. The next terminology test belongs beside those fixtures: retain the audio interval, the intended term and the actual spelling, including an example that receives no attention flag.
Results summarize our October 7, 2026 development record. The real source was an FDA Grand Rounds recording, not patient audio. FAERS and MoCRA are reported as terms encountered in that recording. The full real-audio reference transcript, completed listening adjudication and calibrated terminology detector are not established by these materials.
FAQ
Quick answers to common questions about custom software development, pricing, process and technology.
Type to filter questions and answers. Use Topic to narrow the list.
Showing all 14
No matches
Try a different keyword, change the topic or clear filters
-
How long have you been in business?
Pharos Production has been in business since 2013, with over 13 years of experience in custom software development. During this time, we have delivered over 110 applications for 200+ clients across 28 industries, including FinTech, healthcare, crypto and e-commerce. We are rated 5/5 on Clutch based on 107 verified reviews (2026).
-
What services do you provide?
Pharos Production provides six core service categories: AI Development (LLM applications, AI agents, machine learning and computer vision), Blockchain and Web3 (smart contracts, DeFi, tokenization on Ethereum, Solana, TON and other chains), FinTech Development (payment platforms, digital banking, KYC and AML integration), Software Development (mobile apps, web platforms, database design, UI/UX), Security and QA (code audits, penetration testing, smart contract audits, manual and automated testing) and Consulting (architecture design, MVP validation, startup consulting).
-
Where is Pharos Production located?
Pharos Production is headquartered in Las Vegas, Nevada, USA (5348 Vegas Dr, Las Vegas, NV 89108), with an engineering office in Kyiv, Ukraine (44-B Eugene Konovalets Str. Suite 201, Kyiv 01133). We work with clients worldwide and provide remote collaboration across all time zones. Visit our contact page for directions and scheduling options.
-
How big is your team?
Pharos Production has a team of 90+ engineers, including software developers, blockchain specialists, QA engineers, DevOps experts, UI/UX designers, project managers and solution architects. Our founder, Dr. Dmytro Nasyrov, holds a PhD in Artificial Intelligence and leads the technical direction of all projects.
-
What types of customers do you serve?
We serve a wide range of clients, from startups and product companies to mid-sized enterprises and large institutions. Our clients include crypto exchanges, FinTech providers (like Pleenk), healthcare organizations, sportsbook operators (like Pro Gambling), e-commerce platforms and SaaS companies. Pharos Production has worked with 200+ clients across 28 industries since 2013, adapting engagement models to match each client's stage, whether it is MVP validation for a startup or enterprise-scale development for an established business.
-
How is a custom software development company different from an agency?
A custom software development company like Pharos Production assigns one dedicated engineering team that owns your product across multiple release cycles, instead of handing it off once a contract ends. Agencies more often rotate staff between clients and optimize for fast delivery over long-term product ownership.
Pharos Production keeps the same team on your codebase through 2-week sprints, so there is no re-onboarding cost each time you request a new feature.
-
How much does custom software development cost?
Custom software development costs vary based on project scope and complexity. At Pharos Production, typical project ranges are: MVP development ($10,000-$25,000), suitable for startups validating a product idea; full-fledged production ($25,000-$50,000), for established businesses scaling a proven concept; and full-cycle development ($50,000-$80,000+), for complex enterprise-grade systems. These ranges include architecture design, development, QA testing and deployment. Hourly rates range from $50 to $99 depending on role and seniority. Final pricing depends on technology stack, number of integrations and engagement model (staff augmentation, dedicated team or project outsourcing).
-
How long does it take to develop a custom software application?
Development timelines depend on scope and complexity. At Pharos Production, a typical MVP takes 2-4 months, a production-ready application takes 4-8 months and a complex enterprise system can take 8-12+ months. We use an agile methodology with 2-week sprints, delivering working increments after each sprint. Every sprint includes a retrospective, progress report and planning session for the next iteration. This approach ensures transparency and allows businesses to launch faster by prioritizing high-impact features first. Get a timeline estimate for your project.
-
What industries does Pharos Production serve?
Pharos Production serves 28 industries: FinTech, MiCA Compliance, Banking, Insurance, Crypto, Healthcare, Fitness, E-Commerce, Food & Restaurant, On-Demand, Manufacturing, Automotive, Energy, Oil & Gas, Supply Chain, Construction, Real Estate, Gaming, Streaming, Media, Telecom, Education, Aviation, Travel, Logistics, Marketing, Agriculture and Recommender Systems. Our deepest expertise is in FinTech, blockchain and healthcare, where we have delivered compliance-ready platforms (HIPAA, PCI DSS, GDPR) and high-load systems handling thousands of concurrent users. For the latest industry insights, read our guides on FinTech trends in 2026 and the Web3 technology stack.
-
Why hire a software development company instead of building in-house?
Hiring a software development company offers faster time-to-market, lower upfront costs and access to specialized expertise without long-term employment commitments. According to Deloitte's 2024 Global Outsourcing Survey, 57% of companies outsource software development to access skills they cannot hire internally.
Factor In-house team Software development company Time to assemble 3-6 months (recruiting + onboarding) 1-2 weeks Upfront cost High (salaries, benefits, equipment) Lower (project-based pricing) Specialized expertise Limited to who you can hire locally Access to 90+ engineers across blockchain, AI, FinTech Scalability Slow (each new hire takes months) Fast (scale up or down per sprint) Long-term commitment Full-time employment contracts Flexible engagement models Risk High if key engineers leave Company ensures continuity and knowledge transfer For businesses that need blockchain, AI or high-load architecture expertise, outsourcing to a specialized firm like Pharos Production reduces risk and accelerates delivery.
-
What projects does Pharos Production not take on?
Pharos Production focuses on mid-to-large custom software projects with budgets starting at $10,000. We do not take on template-based websites, WordPress theme customization or short-term contracts under one month. We also do not provide non-technical staffing (marketing, sales or design-only roles). Our strongest fit is blockchain, FinTech and healthcare projects where security, compliance and high-load architecture are critical. For smaller projects or MVPs under $10,000, we recommend exploring freelance platforms or no-code tools as a more cost-effective starting point.
-
Why does Pharos Production use agile methodology?
We use agile with 2-week sprints because it reduces the risk of building features that miss the mark. Each sprint ends with a working demo, a retrospective and a plan for the next iteration.
This means clients see progress every 14 days and can adjust priorities based on real results, not assumptions. We chose this approach after years of experience showing that rigid, fixed-scope contracts lead to scope creep, missed deadlines and products that do not match market needs by launch day.
-
When should you NOT hire a custom software development company?
Custom development is not the right choice in every situation. You should not hire a custom software company if: your problem is fully solved by an existing SaaS product (e.g. Shopify for e-commerce, Salesforce for CRM); your budget is under $10,000 and timeline is under 4 weeks; you need a simple landing page or marketing website (WordPress or Webflow is faster and cheaper); or you are still validating the idea and have not spoken to potential users yet.
In these cases, off-the-shelf tools or no-code platforms offer better ROI. Custom development makes sense when you need unique workflows, regulatory compliance, high-load architecture or competitive differentiation that packaged software cannot provide.
-
What kind of projects has Pharos Production delivered?
Here are three anonymized examples from our recent delivery history:
FinTech startup - payment platform (MVP)
Scope: mobile app + backend API with bank-grade encryption. Team: 4 engineers, 1 QA. Timeline: 10 weeks. Budget: $38,000. Result: launched on schedule, processed $2M+ in transactions within the first quarter.Healthcare provider - patient portal (Full product)
Scope: HIPAA-aligned web platform with EHR integration, appointment scheduling and telemedicine. Team: 6 engineers, 1 DevOps, 2 QA. Timeline: 6 months. Budget: $120,000. Result: 15,000+ active patients, zero compliance violations in two annual audits.Crypto exchange - trading engine (Complex)
Scope: high-load matching engine handling 50,000+ orders per second, multi-chain wallet infrastructure on Ethereum and Solana. Team: 8 engineers, 2 QA, 1 security auditor. Timeline: 11 months. Budget: $340,000. Result: 99.97% uptime, passed three independent security audits.See more projects: NoMoreBets, Pulse, Sagas, Gambit Stream and Pleenk. For the full portfolio, visit our case studies. Learn more about the technology behind these projects in our guide to stablecoins and crypto infrastructure.
Dmytro Ihnatov is the Lead Backend Developer at Pharos Production. He documents backend architecture and research experiments from the team's work on a professional learning platform for aesthetic medicine.
