Engineering library

Systems we build around

Reusable components we apply on client and internal work. Named modules, with a concrete job each. Each one exists because we needed it in production first.

sourcesfilesapisnormalizequery

DataWeave Data

Pipelines and schemas that turn fragmented sources into queryable layers for products and analytics.

Typical jobs

  • Source ingestion from mixed formats
  • Schema normalization and validation
  • Query layers for products and analytics
  • Pipeline health and failure monitoring
textclassifyintententitiessentiment

SentimentCore Language

NLP for intent, entities, and sentiment where text volume requires consistent automated triage.

Typical jobs

  • Intent classification
  • Entity extraction from free text
  • Sentiment scoring
  • Triage routing on model output
historybaselinemodelf(t)

ForecastX Forecasting

Demand, risk, and trend models combining statistical baselines with modern sequence models where they earn their keep.

Typical jobs

  • Demand and volume forecasting
  • Risk scoring over time series
  • Trend and seasonality detection
  • Baseline comparison so gains are provable
streamdetectocranomaly

VisionAI Vision

Object detection, document OCR, and anomaly checks for industrial and operational image streams.

Typical jobs

  • Object and defect detection
  • Document OCR and field extraction
  • Anomaly checks on image streams
  • Batch and streaming processing
rulesscoredecisionapprove

DecisionEngine Automation

Decision flows with explicit business rules, model scores, and human approval steps where required.

Typical jobs

  • Explicit business rule evaluation
  • Model scores as one input among several
  • Human approval steps where stakes require it
  • Audit trail for every decision
trafficchecksallowflag

Sentinel Security

Monitoring for drift, abuse, prompt injection, and data-quality regressions on model-serving paths.

Typical jobs

  • Drift monitoring on live models
  • Abuse and rate-anomaly detection
  • Prompt-injection and manipulation defense
  • Data-quality regression checks
Stack

What we build on

The tools behind the systems above and the platforms we operate. Chosen because we run them in production.

Languages

  • Python
  • TypeScript
  • JavaScript
  • PHP
  • Solidity

Backend & data

  • FastAPI
  • Celery
  • PostgreSQL
  • Redis
  • S3-compatible object storage

Frontend

  • React
  • Vite
  • Vanilla JS & CSS

Infrastructure

  • Docker
  • Caddy
  • Nginx
  • Linux

Observability

  • Prometheus
  • Grafana
  • Structured logging
  • Error tracking

How these get used

These are building blocks, applied inside an engagement. They are applied inside an engagement and shaped to the system they serve. See our services →

Need one of these in your stack?

Tell us what you are running and where it is falling short. We respond from hello@taxeltech.com.