From live market events to cited, trade-ready intelligence — in seconds.
Built from the ground up for financial professionals — not a general-purpose AI adapted for markets. Every layer exists to answer one question: what does a researcher need to know, right now?
Continuous monitoring of regulatory filings, earnings releases, and market events — routed to AI analysis within seconds.
Every AI response is grounded in retrieved source documents, with inline citations on every output.
A live event feed with AI-generated summaries, streamed in real time. One-click deep analysis on any event.
Multi-document Q&A across filings and transcripts, with citations on every claim.
A large-scale model trained on financial data and fine-tuned on proprietary research. Built for finance from the ground up.
All inference and data runs within your private cloud. No shared APIs. Complete data sovereignty.
A market event is detected across all monitored sources within milliseconds of publication.
Relevant documents and market data are retrieved and reranked in parallel.
The model generates structured analysis grounded in retrieved context, with traceable citations.
Analysis streams to the Trader Terminal or Research Workbench with inline citations.
The difference between acting at minute one versus minute thirty is material. Meridian delivers structured analysis before the market moves.
Every response cites its sources. Every figure is traceable.
Your data, your model, your cloud. Nothing flows to shared infrastructure.
Analyst feedback feeds directly into model improvement. The platform gets better with every research cycle.
We work with a select group of financial institutions and professional trading teams. Reach out to discuss your use case.
Meridian processes live market signals simultaneously, synthesising them against a private knowledge base and delivering structured analysis through purpose-built interfaces.
The platform monitors regulatory filing feeds, financial newswires, and market data streams simultaneously — detecting, classifying, and routing events within milliseconds.
A large-scale mixture-of-experts model post-trained on 800 billion financial tokens and fine-tuned on proprietary research examples. Deployed on private GPU infrastructure — no shared inference, no data leakage.
| Specification | Detail | Status |
|---|---|---|
| Architecture | Mixture-of-Experts, 235B parameters | Production |
| Context window | 128K tokens native | Active |
| Time-to-first-token | < 500ms interactive queries | Target |
| Quantisation | AWQ 4-bit — minimal quality loss | Active |
| Domain training | 800B financial token post-training | Complete |
| Fine-tuning cadence | SFT + DPO alignment cycles | Phase 2 |
Two purpose-built interfaces — not generic AI chat adapted for finance. Each is designed around how professional researchers and traders actually work.
Live event feed with streaming AI summaries. One-click deep analysis on any event — streamed to the screen as the model generates it. Built for the first thirty minutes of each trading session.
Multi-document Q&A with inline citations. Regulatory filing browser. Earnings comparison views across periods. Every answer traceable to a specific source document.
All infrastructure runs within your private cloud. No traffic to shared model APIs. No telemetry to external services. Full network isolation enforced at the infrastructure layer.
Every query, response, and data access is logged to a tamper-evident audit store — 7-year retention, queryable for regulatory and compliance review at any time.
An automated classifier screens all prompts and responses for material non-public information indicators — flagging and blocking before sensitive content enters or exits the model.
AES-256-GCM at rest, TLS 1.3 and mutual TLS in transit. Keys managed in a hardware security module. Encryption enforced at the storage, network, and application layers.
We onboard qualified financial institutions through a structured process. Contact us to discuss your use case.
Meridian is the intelligence layer that professional market participants have been waiting for — one that works at the speed of events, grounds every output in real data, and operates entirely within your private environment.
Financial markets generate more structured and unstructured information every day than any research team can synthesise before the market moves.
Existing AI tools were not built for this. They lack domain depth and run on shared infrastructure that institutional security standards cannot accommodate.
Meridian exists to close that gap — delivering grounded, cited intelligence at the speed the market demands, within your private environment.
We never ship an AI output we cannot trace to a source. Citations are not optional — they are the product.
The competitive value of intelligence degrades with time. Every latency target is driven by what matters at a trading terminal at market open.
Our clients' data is not a training set. Private deployment is not an enterprise tier — it is the default architecture for every deployment we do.
We train the model to acknowledge uncertainty as readily as it provides an answer. Trust is earned through consistent accuracy — and honest admission of its limits.
Nogle was founded as a fintech venture builder and investment company, backing and building businesses across the digital financial services ecosystem.
Over nearly a decade, Nogle built and invested across the fintech spectrum — trading platforms, risk management systems, wallet infrastructure, and payment technology. Established offices in Taipei and Singapore. Deployed US$10M in talent and business growth. These years gave us something more valuable than capital: deep, hands-on knowledge of how financial markets operate and where technology falls short.
Drawing on years of operating financial platforms, we began building Meridian — a purpose-built market intelligence platform. Real-time event ingestion, structured data pipelines, regulatory filing monitoring, and multi-asset research infrastructure. A robust, production-grade foundation designed from the outset to carry an AI layer.
Meridian's core platform reaches operational maturity — reliable, institutional-grade, and battle-tested across live market conditions. In parallel, we begin building the AI intelligence layer: assembling the team, domain-training the model, and integrating the components that will transform Meridian from a data platform into an intelligence one.
The AI intelligence layer goes live. Real-time market events become cited, structured analysis in seconds. The platform Nogle spent a decade building toward — private, institutional-grade, and genuinely useful at the speed markets demand.
Expanding Meridian's intelligence across a wider universe of markets, asset classes, and institutions — compounding the advantage of a platform that improves with every research cycle.
Our ambition is straightforward: to make Meridian the platform that serious market participants reach for first — before they read anything else, before they form a view.
We hire experienced practitioners who have solved hard problems in production environments. Meridian runs on live markets, against institutional data, under real latency constraints — and we expect our team to have been there before.
Lead the technical architecture of the Meridian platform. You will set engineering standards, own infrastructure decisions, and work closely with the leadership team on the platform roadmap. Suited to someone equally comfortable in system design discussions and pulling up code.
Apply for this roleOwn the deployment and optimisation of large-scale language models on private GPU infrastructure. Design and operate the retrieval-augmented generation pipeline that grounds every platform output in verified financial data. Production LLM serving experience is essential.
Apply for this roleDesign and maintain the data connectors, event streaming pipelines, and API layer that sit at the core of Meridian. Experience with high-throughput financial data systems and low-latency async architectures is expected.
Apply for this roleArchitect and operate the full data infrastructure layer — vector storage, time-series databases, and object storage — with a strong emphasis on reliability, performance, and infrastructure-as-code discipline. Experience with financial or quantitative data at scale is valued.
Apply for this roleBuild and evolve the Trader Terminal and Research Workbench — data-dense professional interfaces operating on live WebSocket streams and real-time AI output. A strong portfolio of production-grade, performance-critical frontend work is required.
Apply for this roleOwn the Kubernetes-based platform infrastructure, GPU node operations, CI/CD pipelines, and the full observability stack. You will be responsible for the reliability and security of systems that run continuously against live market data. Hands-on GPU workload experience is required.
Apply for this roleA research role at the intersection of financial markets and AI. You will shape how Meridian reasons about market events — defining the analytical frameworks, evaluation standards, and domain knowledge that underpin the platform's intelligence layer. Candidates will bring deep cross-asset research experience from a buy-side, sell-side, or proprietary trading background, combined with a working understanding of how large language models can be applied to financial analysis.
Apply for this roleWe work with systems that financial professionals depend on daily. We look for engineers and researchers who have operated at this standard before and understand what it demands.
Our team takes end-to-end accountability for their domains. We hire people who are comfortable making decisions, defending them with evidence, and iterating when the evidence changes.
Meridian operates within strict security, compliance, and data governance frameworks. We expect our team to treat these as professional requirements, not constraints.
The problems Meridian solves sit at the intersection of financial markets, AI, and real-time systems. We value people who bring genuine depth in at least one of these areas.
Send us your background and the role you're interested in. We review all applications and respond to qualified candidates directly.
We work with financial institutions, professional trading teams, and research organisations. We also respond to every career enquiry personally within 72 hours.
We onboard qualified institutions through a structured process. Expect a response within one business day of your enquiry.
All 7 founding roles are open. Include the role you're applying for and attach or link to any relevant work.