Acyclic Research is a cooperative network of advanced applied mathematicians and experienced developing markets fixed income professionals. We provide authoritative risk analysis for some of the world's most information-denied regions.
Our approach is quantitative, using the foundational components of machine intelligence. However the complex developing markets domain can be out-of-band for mainstream off-the-shelf AI services. This is when our global network of relationships, which reaches deep into the Global South, and our ability to handle ragged data from challenging sources, come into play to augment purely online methods.
The fusion of nuanced human experience, custom data, and advanced models, provides confidence for even the most complex risk assessment mandates.
Services
Here is a sample of the projects we have already undertaken for our partners.
Semantic Trend Analysis
AI-enabled extraction of thematic scores from unstructured text across frontier market sources.
Primary Due Diligence
Deep background pre-issuance research combining digital and off-grid sources in opaque jurisdictions.
Real-time Anomalies
Dimensionality reduction of tens of thousands of live data series, for early detection of emerging risks in hidden factors.
Complex Pricing
Deterministic and stochastic modelling of bonds, loans, and structured products in high-entropy environments.
Custom Surveys
Interviews and surveys of risk-influential networks for proprietary intelligence on specialised topics.
AI engineering
Parsing and homegenisation of custom data sources; design and implemention of AI processing pipelines.
Physical Communication
In an era of generative digital oversupply, attention bandwidth is narrower than ever. We attach high importance to incisive and elegant outputs for all our analyses. Our team's many years of front-office experience allows us to complement online channels with engaging real-world communication, including publication-quality print media, event organisation, targeted roadshows, seminars, and presentations.