Skip to content
BERNARD ABUTO

02Work

Work at the intersection of research and systems.

These projects explore different parts of the same problem: how to turn technical ideas into systems that can be understood, evaluated, and used. Some are active engineering efforts; others are ongoing research. Each reflects a different part of how I approach building.

P.01African Speech & Language AIResearch & Products Company

Sauti Labs

An AI research and products company building the models and technologies that let machines understand and generate speech across Africa’s linguistic diversity.

Fig. A — Research to productIllustrative

Research teams

  1. 01Speech IntelligenceRecognition, synthesis, voice
  2. 02Language IntelligenceUnderstanding, generation, evaluation
  3. 03Language ResourcesDatasets, corpora, provenance
  4. 04Model IntelligenceTraining, adaptation, efficiency
  5. 05Responsible AISafety, privacy, misuse prevention
Sauti CloudProduct · Early access
  • Speech recognition
  • Speech generation
  • Voice AI
How Sauti Labs’ research teams feed its product. Sauti Cloud is in development with early access; it is not yet generally available.
P.02AI Systems & Offline-First ApplicationsActive Engineering

Scribe AI

An offline-first church production system combining a local speech-intelligence prototype with a separately developed desktop application architecture.

Fig. B — Scripture pathIllustrative
  1. Audio input
  2. Transcription
  3. Scripture reference detection
  4. Local Bible lookup
  5. Human approval
  6. Display
Not in live pathService-state machine · Lyric detection
The solid path is the scripture workflow connected in the live entry point. The dashed modules exist and are tested, but are not yet integrated into the live audio path.
P.03Product Engineering & Software SystemsIn Development

Rentra

A rental operations platform modelling the relationships, records, and workflows connecting tenants, landlords, properties, and rental activity.

Fig. C — System & domain structureIllustrative
  • Tenant
  • Landlord
  • Caretaker
  1. Flutter mobile appClient
  2. NestJS serviceBackend
  3. PostgreSQL · SupabaseData
Drawn from the repository’s architecture documentation. This is a structural diagram, not a screenshot of the application.
P.04Applied Machine Learning ResearchActive Research

QueleaGuard

Investigating environmental, spatial, temporal, and agricultural drivers of red-billed quelea occurrence around Ahero Rice Irrigation Scheme.

Fig. D — Study designIllustrative
Schematic analysis grid of roughly 5.5 kilometre cells within a 50 kilometre extent around Ahero Rice Irrigation Scheme, with Lake Victoria to the west and a river crossing the grid.AHERO SCHEMELAKE VICTORIA(WINAM GULF)RIVER≈ 50 KM EXTENT1 CELL ≈ 5.5 KMN
Schematic, not to scale. The grid, lake, and river are drawn to explain the study design; no observations, predictions, or model outputs are shown.