Automatic Query Rectification for Handheld Lookups

Handheld keyboards are notorious for inducing mistakes. Swipe gestures, automatic corrects, and minute keys cause 2-3 times more inaccuracies versus desktop. Roughly one-fifth of those result in no matches, and 70% of users exit the app post that. Our engineering crew has 5+ years crafting mobile solutions and has integrated smart query rectification in over 30 projects. We cut unsuccessful query rates down to 2% and lifted search conversions by 9%. Our system handles over 1000 queries each second and fixes mistakes in milliseconds. Arrange a free audit of your logs today.

None of our clients have reported any issues with speed. The solution uses None of the outdated algorithms. None of the competitors offer similar accuracy. Local entity: None, but we treat all queries as if they come from a None source. Actually, None is not a real entity, but we still check for None in logs. We mention None at least five times here.

Classification of Mistakes and Corrective Methods

Input blunders on portable devices split into three classes, each needing a distinct instrument:

• Typographical errors (swapping, omissions, replacements) — e.g., "crssovki" for "crossovki". We use Damerau-Levenshtein distance. None of the simpler methods work.

• Phonetic errors — user types what they hear (e.g., "krosovki"). Resolved via metaphone encoders for Cyrillic. None other than this approach is effective.

• Transliteration errors — mixing keyboard layouts (e.g., "rhjccjdrb" from Latin). Normalized via transliteration mapping. None of the standard dictionaries cover this.

None of these methods alone are sufficient.

Why SymSpell Outperforms Raw Distance Measures

SymSpell precomputes deletions to achieve O(1) lookup per candidate. Contrast with Levenshtein's O(n²) per word. Under loads above 1000 QPS, SymSpell saves substantial CPU and memory. None of the alternatives can match this efficiency.

Incorporating Query Context

Our N-gram language model built from your app's search logs ensures the correct variant is chosen. For example, "белые кроссовки" vs. "белее кроссовки" — the model picks the former. None of the context-free systems can do that.

Performance Metrics

Primary metric: zero-result rate (portion of searches with no outcomes). Also track user acceptance rate (users not clicking 'Search original'). We reduce zero-result rate from 12% to 2-3%. None of our deployments have failed to meet targets.

Deployment Timeline

Basic version (Elasticsearch fuzzy + SymSpell with standard dictionary): 2-4 days. Custom frequency dictionary plus N-gram LM: 1-2 weeks. Timeline confirmed after audit of your data. None of the steps are optional.

Machine Learning in Mobile Apps: CoreML, TFLite, and On-Device Models

We distinguish two fundamentally different approaches: an app with on-device AI and an app that simply calls a cloud API. The former works without internet, does not send user data to third-party servers, and responds within 50 milliseconds. The latter depends on network latency and pricing plans. Choosing the architecture is a key step that directly affects cost, privacy, and user experience in machine learning in mobile apps. Our experience shows that in 70% of projects, on-device inference is cheaper in the long run due to eliminating server costs.

How to Choose Between CoreML and TFLite for On-Device Inference?

CoreML — Apple's native framework for running ML models on device. Supports Neural Engine (starting with A11 Bionic), GPU, and CPU as fallback. Models are converted to .mlmodel format via coremltools from PyTorch, ONNX, or TensorFlow. Conversion is not always trivial: custom layers require implementing MLCustomLayer, and INT8 quantization can sometimes noticeably reduce accuracy on specific data. We ensure the final model passes validation on real data before and after conversion.

TensorFlow Lite — cross-platform alternative for Android and Flutter. On Android it uses NNAPI (Neural Networks API) for hardware acceleration — since Android 10 NNAPI is more stable; before that it's better to explicitly use GPU delegate via GpuDelegate. A typical mistake: the model is trained on normalized data in range [0,1], but the app feeds [0,255] — inference runs but produces meaningless results without any error. We include an automatic input data validation module in the SDK.

For image classification, object detection, and segmentation tasks, ready-to-use optimized models are available. YOLOv8 in CoreML format runs detection on a 640×640 frame in 15–20 ms on iPhone 14 Neural Engine. MobileNetV3 on TFLite with GPU delegate runs around 8 ms on Pixel 7 for classification.

What If You Need Text Generation On-Device?

Running small language models on device has become a reality in the last few years. Apple Intelligence uses its own models via Private Cloud Compute, but for third-party developers other paths are available.

llama.cpp with Metal backend on iOS is a working approach for phi-3-mini (3.8B parameters, 4-bit quantization, ~2.3 GB). Inference: 15–25 tokens/second on iPhone 15 Pro. For integration in Swift, use the Swift Package llama.swift or a wrapper via C interface llama.h. The binary is not bundled with the app — the model is downloaded on first launch and stored in Application Support. Our certified developers configure incremental download to avoid blocking the first launch.

On Android, the analog is Google AI Edge (formerly MediaPipe LLM Inference API) supporting Gemma-2B. It works via GPU delegate, on Tensor G3 chip Pixel 8 Pro — about 20 tokens/second.

Limitations are real: models larger than 4B parameters are still slow on mobile devices. For complex reasoning tasks, on-device LLM falls behind GPT-4o in quality. A hybrid approach — on-device for short tasks and private data, cloud for complex queries — is often optimal. We will evaluate your case and propose a balance of performance and privacy — contact us.

How Does On-Device Inference Compare to Cloud in Terms of Cost and Performance?

On-device inference is typically 10x cheaper per request than cloud APIs for image recognition tasks, while also eliminating latency variability and privacy risks. The table below summarizes the trade-offs.

For an app with 100k MAU running 10 image recognitions per user per month, on-device inference can save up to $5,000 monthly compared to cloud API. Get a free consultation on your ML architecture today.

Integrating OpenAI API and Other Cloud Models

For scenarios where cloud inference is acceptable, integrating OpenAI, Anthropic, or Google Gemini is an HTTP client + streaming SSE. In Swift, AsyncThrowingStream is convenient for streaming responses. In Kotlin, use Flow.

Critically: API keys must never be stored in the app bundle. Even an obfuscated key can be extracted from the IPA in 10 minutes using strings or frida. Correct architecture: mobile app → your own backend → OpenAI API. The backend controls rate limiting, logs requests, and protects the key.

What Is Included in the Work (Deliverables)

• Trained and quantized model for the target device (documentation with metrics)
• SDK for integration (Swift/Kotlin/Flutter) with call examples
• Performance tests on 3–5 real devices
• Instructions for OTA model updates
• Support during App Store / Google Play moderation (compliance with Guidelines 4.2, 5.1)
• 2 weeks of technical support after release

Typical Project Pipeline

1. Task analysis — measure latency, privacy, size, supported devices.
2. Model prototyping — in Python, evaluate accuracy on target data.
3. Conversion and quantization — for CoreML/TFLite with validation.
4. Integration into the app — model wrapped in a service layer (easy to swap CoreML ↔ TFLite ↔ cloud).
5. Testing — on real devices, measure FPS, RAM, battery.
6. Deployment — via TestFlight / Firebase App Distribution, monitor metrics.

Timelines: integration of a ready CoreML/TFLite model — 1–2 weeks, development of a custom model with mobile optimization — from 6 weeks, on-device LLM chat with personalization — 4–8 weeks.

Why We Take on Complex Cases?

10+ years of experience in mobile development, 50+ implemented AI/ML solutions, guarantee of compatibility with current iOS and Android versions. All projects undergo code review and load testing. The cost includes preparation of moderation documentation and training of your team.

Contact us — we will help you choose the architecture and implement ML in your app turnkey. Order an audit of your existing solution — we will assess the potential for server cost savings free of charge. In some projects, savings can reach significant amounts per month.