public analytical areas
Recorded public product scope.
← All work
Market intelligenceResearch prototype
Cakrawala Intelligence Terminal
Forex-first market and macro research that makes source health, evidence freshness, model state, and decision boundaries visible.
Market intelligence2026-08-29
Cakrawala Intelligence Terminal
12public analytical areas
Dailyresearch cadence
Weeklymodel governance
Quick read
Project summary
Problem
A market conclusion is weak when the evidence trail is hidden. Provider reliability, point-in-time evidence, authorization, and model state have to be treated as product requirements.
My role
I designed the research terminal, provider and evidence boundaries, model-governance rules, and the separation between public research and owner-only workflows in the reviewed scope.
Result
Recorded product scope covers 12 analytical areas with daily research and weekly model-governance design.
Stack
Python · Streamlit · market data adapters · model governance
Status
Research prototype
research cadence
Market state and event-risk analysis.
model governance
Review does not force model replacement.
decision rule
AI explains evidence rather than inventing signals.
Overview
Cakrawala is a market and macro research terminal designed around evidence quality, freshness, model state, and explicit decision boundaries.
Problem
A market conclusion is weak when the evidence trail is hidden. Provider reliability, point-in-time evidence, authorization, and model state have to be treated as product requirements.
Why it matters
Market conclusions are easier to challenge when source health, evidence freshness, model state, and authorization boundaries are visible instead of being hidden behind a single narrative.
My role
I designed the research terminal, provider and evidence boundaries, model-governance rules, and the separation between public research and owner-only workflows in the reviewed scope.
Product
The public product spans 12 analytical areas and keeps public research separate from owner-only workflows.
- Market structure, macro context, positioning, event risk, and risk tools are presented as research evidence.
- AI is constrained to explanation and research assistance instead of creating decisions by itself.
- Time-sensitive market conclusions remain in the live product rather than being frozen into static portfolio claims.
Data
The design requires external providers to prove accessibility, attribution, usage rights, schema quality, freshness, fallback behavior, and network safety before their evidence is trusted.
- Missing or unhealthy evidence fails closed instead of being silently accepted.
- Source health and evidence freshness are intended to be visible product states.
Method
Market analysis and model governance are deliberately separated from language-model explanation.
- Daily market-state and event-risk analysis.
- Weekly model review that does not force automatic replacement.
- Champion promotion requires configured gates and out-of-sample evidence.
- Deterministic decision policy remains separate from AI explanation.
Results
The current portfolio evidence is product scope and governance behavior rather than a claimed trading return.
- 12 recorded public analytical areas.
- Daily research cadence and weekly model governance in the reviewed product description.
- Public and owner-only boundaries are explicitly separated.
Validation
The system is designed to reject weak or stale evidence rather than fill gaps with confidence.
- Provider checks cover accessibility, attribution, usage rights, schema, freshness, fallback, and network behavior.
- Champion model promotion requires out-of-sample evidence and configured gates.
- Personal Mode authorization is server-side rather than a hidden frontend state.
Architecture
The architecture separates provider evidence, research logic, deterministic decisions, AI explanation, and authorization boundaries.
- Provider registry and source-health layer.
- Point-in-time evidence envelopes and research state.
- Deterministic decision policy and risk gates.
- AI explanation layer that does not own the trading decision.
- Provider registryApproved market and macro sources
- Source healthFreshness, schema, and fallback checks
- Research statePoint-in-time evidence envelopes
- Decision and risk gatesDeterministic policy before explanation
- Research interfacePublic insight and owner-only boundaries
Tradeoffs
The project favors visible source quality and explicit failure states over broad provider coverage or constant model replacement.
Limitations
The portfolio does not claim that the research terminal guarantees profitable decisions. Source migration is still in progress, so a remote repository is not presented as authoritative evidence.
Live Product
A public Streamlit URL is recorded in prior project evidence, but direct availability could not be independently reverified in this environment. No recruiter-facing live link is published until that check succeeds.
Source
Source migration is still in progress. This case study therefore describes verified product behavior and recorded design boundaries without claiming that an unverified repository is the source of truth.
Technical status: Recorded public product, direct availability unverified
Repository status: Source migration is in progress, so no repository is presented as authoritative here.
Next milestone
Complete provider-registry validation, immutable evidence envelopes, source-health persistence, production-quality market adapters, and fixture-based integration tests.