AI System for Formulation Optimization and Product Composition

When developing new formulations, R&D departments face a multi-parameter challenge: simultaneously improve functional properties, reduce cost, meet regulatory limits, and maintain sensory characteristics. Experimental trial-and-error means weeks of lab work and hundreds of thousands in costs. We offer a different path: an AI system that finds optimal compositions in 3–5 months, saving up to 30% in cost without sacrificing quality. Our experience in ML optimization spans 7 years and 40+ completed projects.

A typical scene: a technologist manually adjusts the composition, runs tests, adjusts again. One iteration takes a day or two. For a complex formulation, 20–30 iterations are needed — a month of work and significant raw material costs. Moreover, the found solution is often not optimal — it merely meets minimum requirements. AI optimization cuts the number of experiments by 3–5 times and guarantees finding a composition as close to the ideal as possible according to given criteria.

Application Areas

• Food industry: product compositions (taste/nutrition/cost balance)
• Cosmetics: cream, shampoo formulas (efficacy/safety/sensory)
• Pharmaceuticals: solid dosage forms, excipients
• Chemicals: coatings, adhesives, polymers, fuels
• Construction materials: concrete, coatings, sealants

How AI Reduces Formulation Cost?

The traditional path is trial and error: a technologist selects components, tests, adjusts. This is expensive and slow. The AI approach uses surrogate models — ML models that predict formulation properties based on composition. For example, XGBoost or Gaussian Process are trained on historical data and guide Bayesian Optimization toward the optimal region. Within 50–150 iterations, the algorithm finds an ingredient combination that meets all constraints at minimum cost.

Mathematical Formulation

Variables: proportion of each ingredient in the formulation (x₁, x₂, ..., xₙ)

Constraints:

• Σxᵢ = 1 (sum of proportions = 100%)
• 0 ≤ xᵢ ≤ max_i (maximum content limits for each component)
• Regulatory: xₐ ≤ 0.1% (food additives), xᵦ ≥ 0.5% (minimum ingredient)
• Technological: ingredient compatibility

Objective functions (multi-objective):

• Minimize cost
• Maximize functional properties (strength, viscosity, pH)
• Meet sensory characteristics (taste, consistency)
• Minimize undesirable components

ML Methods for Formulation Optimization

Surrogate Models

Problem: physical testing of a formulation is expensive and slow. A surrogate ML model predicts composition properties from its components:

• XGBoost/Random Forest on experimental data
• Gaussian Process (with uncertainty quantification — we know where we are unsure)
• Deep Learning for complex structural dependencies

Bayesian Optimization

Active cycle: surrogate → acquisition function → next experiment:

1. Initial DoE: 20–50 formulations with uniform coverage of the space
2. Train GP surrogate
3. Expected Improvement selects the next most informative point
4. Experiment → update surrogate
5. Convergence to optimum in 50–150 iterations

from bayes_opt import BayesianOptimization
import numpy as np

def formulation_score(ingredient_a, ingredient_b, ingredient_c, ingredient_d):
    """Surrogate model: prediction of target properties"""
    # Normalization of ingredient sum
    total = ingredient_a + ingredient_b + ingredient_c + ingredient_d
    fractions = [x/total for x in [ingredient_a, ingredient_b, ingredient_c, ingredient_d]]

    # ML surrogate (pretrained model)
    features = np.array(fractions).reshape(1, -1)
    predicted_quality = surrogate_model.predict(features)[0]
    predicted_cost = cost_model.predict(features)[0]

    # Multi-objective scalarization
    score = predicted_quality - 0.3 * predicted_cost
    return float(score)

optimizer = BayesianOptimization(
    f=formulation_score,
    pbounds={
        'ingredient_a': (0.05, 0.60),
        'ingredient_b': (0.10, 0.45),
        'ingredient_c': (0.05, 0.30),
        'ingredient_d': (0.02, 0.20),
    },
    random_state=42
)
optimizer.maximize(init_points=20, n_iter=100)

Mixture Design + ML

Specialized Design of Experiments for mixtures: Simplex-Lattice, Simplex-Centroid, D-Optimal designs. Combination of statistical DoE with ML modeling.

Why Bayesian Optimization is More Efficient Than Standard DoE?

DoE (Design of Experiments) gives good space coverage but does not adapt to data. Bayesian Optimization uses results from previous experiments to guide the search toward the region where improvement is expected. This cuts the number of required tests by 3–5 times. Additionally, Gaussian Process provides a measure of uncertainty — you know where the model is unsure and can run an additional experiment exactly there.

Robust Optimization

Ingredients from different batches have different characteristics. Robust optimization includes:

• Monte Carlo simulation of input component variability
• Robust formulation: performs well under ±σ variation of ingredients

Formula Knowledge Base

NLP search across historical formulation database: "find formulas with viscosity 1000–2000 cP and preservative content <0.5%". Semantic search on chemical descriptions. Substitution engine: ingredient discontinued → ML proposes an alternative that preserves key properties.

What Does System Development Include?

1. Data analysis: collection of historical formulations, test results, constraints.
2. Surrogate model building: training ML models to predict properties.
3. Bayesian Optimization implementation: tuning acquisition function, multi-objective handling.
4. Integration with LIMS: automatic data transfer from the laboratory information system.
5. Validation: testing on a holdout set, comparison against baseline.
6. Documentation and training: instructions for technologists, handover of the model to production.
7. 6-month support: model adjustments as more data accumulates.

Work Process

Total timeline: 3 to 5 months. Exact estimate after audit of data and constraints.

Typical Mistakes in Implementing AI Formulation Optimization

• Ignoring raw material variability: a model trained on one batch fails on another.
• Too few initial data: Bayesian Optimization needs at least 20–50 experiments to start.
• Ignoring multi-objective: optimizing only cost leads to quality loss.
• Lack of human oversight: AI suggests optimal composition, but technologist must confirm.

We guarantee the system will account for all technological and regulatory constraints. We provide model correctness certificates based on validation results.

Want to evaluate the potential of AI optimization for your formulations? Contact us — we will conduct a data audit and provide a preliminary estimate. Get a free consultation with an AI/ML engineer.

Industry AI Solutions: Healthcare, Finance, Retail, Manufacturing

We encounter the same pain points: a general text model doesn’t distinguish medical nomenclature, and a standard object detector confuses “weld seam scratch” with “casing scratch.” Each time these are different defects with different consequences. To avoid this, we build industry-specific solutions on top of general methods, but with deep domain knowledge — from regulatory requirements to data specifics. Over 5 years, we have completed 80+ projects in fintech, healthcare, retail, and manufacturing, and none were without adaptation to a specific business case.

Healthcare: Regulatory Maze and Data Governance

Medical AI differs not in technical algorithms but in a compliance-first approach. Depending on the country of application, the model may be a Class II or III medical device requiring clinical trials (FDA, CE MDR, GOST R). We ensure compliance with these standards at the architecture stage — fixing them post-factum is 10× more expensive.

Medical imaging. Detection on X‑rays, CT, MRI is a mature area. Models on ResNet, EfficientNet, SegFormer achieve AUC 0.94–0.97 on standard tasks (pneumonia on CXR, polyps on colonoscopy). Key issue is generalization: a model trained on data from one scanner manufacturer degrades on another due to differences in preprocessing and artifacts. Solution: domain adaptation via MONAI (Medical Open Network for AI) from NVIDIA, which includes DICOM loading, 3D augmentation, and confidence calibration. TotalSegmentator — for automatic segmentation of 117 structures on CT, production‑ready, Apache 2.0 license.

Clinical NLP. Extracting structured information from clinical records: diagnoses (ICD‑10/11), prescriptions, dates, indicators. medspaCy, scispaCy, MedCAT — specialized NLP libraries with ontologies (SNOMED‑CT, UMLS). Fine‑tuning BioBERT or ClinicalBERT on our data yields F1 0.85–0.92 on NER tasks versus F1 0.65–0.72 for general BERT. We verified this on a project with a regional oncology center — cancer stage extraction accuracy increased by 23%.

Clinical decision support. LLM assistants for clinical decision support are a regulatory gray area. We use an RAG system on top of clinical guidelines (UpToDate, local protocols) with explicit citation for each statement. The model does not diagnose but helps find relevant protocols. Stack: LlamaIndex + pgvector + pubmedbert-base-embeddings + Llama Guard for safety. Data in DICOM/HL7 FHIR, on‑premise deployment mandatory.

Finance: How to Ensure Interpretability of a Scoring Model under Basel IV?

The financial sector is one of the most mature in applying ML, but regulation is maximal. Every model affecting credit decisions falls under Basel IV, EU AI Act, GDPR Article 22. We deliver AI solutions for fintech that satisfy these requirements — in a project for a top‑10 bank we deployed a scoring model where each record required SHAP explanations.

Credit scoring. Gradient boosting (LightGBM, XGBoost) dominates. Neural networks yield +0.5–2% AUC but lose interpretability. Standard: LightGBM + SHAP to explain each decision. Fairness checking is mandatory: Fairlearn or aif360 for auditing disparate impact on protected attributes (age, gender). The default class is 1–5% — with an imbalance of 1:30, a model with 97% accuracy may have recall 0.2. Solution: focal loss, class_weight='balanced', SMOTE + careful validation. In one fintech scoring project, the model reduced credit losses by $2.1 million annually.

Algorithmic trading and risk management. LSTM and Transformer for price forecasting are popular but unstable in production due to non‑stationarity of financial series. A more robust approach: ML for signal generation (classification: up/down over horizon N) with traditional portfolio optimization on top. Backtesting via Zipline‑Reloaded, vectorbt, QuantLib. Proper backtesting is critical — look‑ahead bias kills results. We guarantee a clean experiment: all data at signal time is available in real time.

AML (Anti‑Money Laundering). Graph Neural Networks for analyzing transaction networks is an actively developing area. PyG, DGL for GNN. Task: detect suspicious patterns in transaction graphs (layering, structuring). Recall is more critical than precision — better 10 false alarms than miss one money laundering. In a project for a large payment service, we increased recall by 18% without increasing false positive rate.

Retail and e‑commerce: Recommendation Systems and Demand Forecasting

Recommendation systems. Current architectural standard: two‑tower model for retrieval + ranking with cross‑features. TensorFlow Recommenders or Merlin from NVIDIA for GPU‑accelerated feature processing. For small catalogs (<100k items), LightFM is sufficient. A common mistake is training on implicit feedback without accounting for position bias. Solution: IPW (Inverse Propensity Weighting) or randomized logging on a portion of traffic. Development time for a basic recommendation system is 4–8 weeks, including A/B test.

Demand forecasting and inventory optimization. Hierarchical forecasting: SKU → category → store → region. HierarchicalForecast from Nixtla automatically reconciles forecasts across levels. TFT or N‑HiTS for base forecast, gradient boosting for adjustment on exogenous factors (promotions, weather, events). One retail project led to a 15% reduction in stock‑outs due to precise promotion calibration.

Visual search and size compatibility. CLIP embeddings for image search — deploy in 2–3 weeks: clip‑ViT‑B‑32 or clip‑ViT‑L‑14, Faiss or Qdrant index, REST API. For size recommendation — specific models on return data and reviews with fit indication.

Manufacturing: Quality Inspection and Predictive Maintenance

Quality control and defect detection. CV models for product inspection are one of the most mature industry tasks. YOLOv10 for defect detection, SegFormer for segmentation. Specifics: class imbalance (defects are rare), high recall requirement (missing a defect is worse than false alarm). Typical dataset: 500–2000 defect images + 500–1000 normal. Few‑shot learning via DINO or SAM 2 works with 50–100 annotated examples. We gained experience on an electronics production line — recall 0.95 at FPR 0.03. A predictive maintenance deployment saved a manufacturing client $500,000 per year in unplanned downtime.

Predictive maintenance. Vibration sensors, current sensors, thermocouples → feature extraction → anomaly or mode classification. Models: LSTM‑AE for unsupervised, LightGBM for supervised (if failure history is available). Integration with SCADA/OPC‑UA via opcua-asyncio or MQTT. Key metric: False Negative Rate — a missed pre‑failure is more costly than a false alarm. Threshold tuned to business cost of each error type. Timeline: 3 to 6 months to production.

Digital twin and simulation. Surrogate models — ML models replacing expensive physical simulation. If a CFD simulation takes 6 hours and a surrogate (trained on 10,000 simulations) takes 0.01 seconds, that's 2,000,000× speedup for optimization. SALib for sensitivity analysis, botorch for Bayesian optimization on top of surrogate.

General Principles of Industry AI

Regardless of industry, there are patterns that work everywhere. Data matters more than architecture. In healthcare, 1000 quality labeled images are better than 100,000 poor ones. In manufacturing, 200 real defect examples are more valuable than 10,000 synthetic ones. Compliance‑first design — regulatory requirements are easier to embed into architecture from the start than to add later. Logging, explainability, versioning from day one. Domain expert on the team — an ML engineer without domain knowledge does slowly and error‑prone what an ML engineer plus a doctor/financier/technologist does quickly and correctly.

We guarantee certification to customer requirements (ISO 13485, SOC 2, GDPR) and provide full model documentation (model card, datasheet, compliance report). Our experience: 10,000+ engineering hours and 80+ projects.

Work Process for an Industry AI Solution

1. Domain immersion (2–3 days) — interviews with experts, studying regulatory requirements, auditing available data.
2. MVP design (1–2 weeks) — stack and architecture selection, feasibility assessment.
3. Development and validation (from 4 weeks to 6 months depending on industry) — model training, testing, compliance.
4. Integration and deployment (1–4 weeks) — on‑premise / cloud / edge, documentation, staff training.
5. Support and monitoring — model drift, retraining, SLA.

Estimated timelines:

Cost is calculated individually for each project. Get a consultation — we will evaluate your dataset, regulatory map, and business goals.

Why Choose Our Industry AI Solutions?

• 80+ completed projects in fintech, healthcare, retail, and manufacturing.
• 5 years on the market — proven experience with compliance and deployment.
• Quality guarantee: we ensure target metrics (AUC, recall, latency p99) and provide full documentation.
• Licensed technologies: PyTorch, MONAI, LightGBM, Qdrant — we use open‑source with commercially safe licenses.
• Flexibility: we work as a contractor or as an extension of your team.

Contact us for a free data audit and consultation. Request a proposal with a detailed work plan. We will discuss your task and prepare a commercial proposal.