Radix Analytics Pvt Ltd

Radix Analytics is a custom data science firm headquartered in Singapore with over a decade of experience. The team is led by domain-specialist PhDs and combines data scientists, data engineers, and software developers. They've completed 50+ projects across five continents and serve clients in Australia, the UK, USA, Singapore, Dubai, India, Canada, Vietnam, Malaysia, and Indonesia. Their core strength is turning raw data into production-grade decision systems through machine learning, statistical models, and mathematical optimization delivered as APIs.

Services and capabilities

Radix operates across three core pillars: data analytics, AI/ML, and data engineering.

In data analytics, they build statistical models for pricing optimization, inventory planning, risk scoring, and demand forecasting. They also ship executive dashboards with smart alerts and integrate solutions into existing systems.

For AI/ML, Radix builds supervised and unsupervised models for forecasting, anomaly detection, and customer segmentation. They handle NLP, computer vision, and deep learning. On the generative AI side, they offer chatbots that answer questions over documents with citations, access-rule enforcement, and on-premises options. They include lightweight model adaptation (LoRA/PEFT), preference optimization (DPO/RLHF), and agent workflows that call APIs. Everything ships with MLOps: experiment tracking, model registry, CI/CD, canary/blue-green rollouts, performance monitoring, and drift detection.

Data engineering work covers connecting apps, files, APIs, and databases; data cleaning, deduplication, and modeling; monitored batch and stream pipelines with SLAs; and layered data lakes and warehouses with semantic layers, versioning, and secure APIs.

They emphasize explainability in models—shipping model cards with metrics, bias/variance notes, and operating thresholds alongside inference APIs with SLOs and runbooks for retraining and incident response.

Industry-specific delivery

Radix has built specialized practices in five sectors.

Financial Services: Credit risk analytics and ML scorecards for banks, auto financiers, leasing firms, SME lenders, microfinance, and affordable-housing providers. They deliver reason codes and decision overlays to reduce discrimination and align underwriting with scores. For SME lenders, they use thin-file and alternative-data modeling. For auto loan providers, application scorecard models align rules and reporting to reduce manual scrutiny. In microfinance, they build lightweight, high-recall models for first-time borrowers.

Supply Chain & Retail: Forecasting, inventory science, and planning software for retail, distribution, shipping, ports, and industrials. They deliver demand signals, right-sized policies, and real-time visibility across the chain.

Media: They sell two products—Ad-Spot Optimizer (ASO) for real-time spot allocation to maximize revenue without breaking deal commitments, and Ad-Revenue Optimizer (ARO) for dynamic pricing and advertiser-aware proposals. They also offer AI-assisted content scheduling to maximize reach while controlling viewer fatigue. Their media clients have seen 5–10% revenue margin uplift and 2–5% revenue uplift in production.

Industrial, Energy & Natural Resources: Decision systems for manufacturing, energy, oil & gas, and mining focused on asset utilization, long-horizon planning, inventory sizing, scheduling, and operational safety.

Public Sector & Education: Category planning, store operations, civic systems, and learning platforms with a focus on measurable outcomes—fairness, transparency, and real-world value.

How they work

Radix uses a five-step methodology: Discover (map decisions, constraints, and data fidelity), Diagnose (quantify value gaps using exploratory analysis), Design (prototype on real data with features, rules, and simulations), Deploy (production APIs and dashboards with monitoring, integrated with existing systems), and Derive Value (set KPIs, run A/B and champion-challenger tests, institutionalize winning policies).

They work as an embedded team, as evidenced by client feedback noting seamless integration into existing product and engineering teams. Engagement appears flexible across different problem types—from pure IT to complex data analytics.

Notable work

For Address/Entity Matching, a client needed ML to find relations and standardize messy data. Radix built a two-stage ML approach that plugged into existing systems via APIs and scaled to large databases. It achieved 87% recall for addresses and 89% for entities.

For an edTech test-prep company, Radix solved a complex text- and image-based classification problem involving MCQ questions with mathematical symbols, structures, equations, and diagrams. They delivered 92% accuracy across about 30 topics in each subject.

For a major Singapore insurance company, Radix deployed a Fraud Scoring Tool for travel insurance. The model handled low fraud-incident rates and factors like claim nature and police complaint status. It reduced manual investigation by 60% while maintaining high recall and substantial precision.

Other named clients include FMart (inventory planning and demand forecasting), Pedagogy (product transformation), Blue Ocean Holiday Homes (booking platform with administration, agent, client, and front-desk modules), Corpository (IT and analytics), and Agrim HFC (credit and analytics).

Team and credentials

The firm is led by domain-specialist PhDs and includes skilled data scientists, data engineers, and software developers. No team size or specific certifications are mentioned in the source material.