Research infrastructure for markets that refuse to behave simply.
NRGX Labs is a founder-built market research platform that brings independent quantitative engines, options intelligence, volatility analysis, market structure, and AI-assisted synthesis into one connected system.
18 analytical engines ·
190+ API routes ·
160,000+ lines of Python and JavaScript
One continuously evolving research environment.
No single market model deserves complete trust.
Market conditions are nonlinear. The same price path can carry very different meanings depending on volatility, positioning, liquidity, and the macro calendar behind it. A single indicator, however sophisticated, creates blind spots.
Options data regularly contradicts spot-price narratives. Implied volatility, skew, and open interest often reprice risk hours or days before the underlying confirms it.
Volatility, dealer positioning, flow, macro conditions, and market structure interact. Examined in isolation, each one tells a partial and sometimes misleading story.
Regimes change faster than models retrain. A signal that worked in a compressed volatility environment can invert quietly when the regime shifts underneath it.
One master signal concentrates model risk. Independent engines that can agree, disagree, and abstain distribute it, and disagreement itself becomes information.
NRGX Labs is built around multiple independent engines rather than one master signal. When the engines disagree, that disagreement is the finding.
A research operating system, not a dashboard with charts.
The platform is organized as connected layers. Each layer observes, transforms, scores, or synthesizes a specific slice of market information, and every layer is inspectable from the desk.
Market Structure
Tracks index and single-name structure: expected move, open-interest clusters, support and resistance formed by positioning rather than price memory.
Options and Volatility
Builds volatility surfaces, monitors skew and term structure, and measures how implied risk pricing changes across strikes and expirations.
Flow and Positioning
Reads dealer gamma context, open-interest migration, and flow concentration to estimate where hedging pressure amplifies or dampens moves.
Regime Detection
Classifies the prevailing volatility and trend regime, and flags transitions where historical relationships are most likely to break.
Risk and Scenario Analysis
Runs Monte Carlo simulation, cross-asset stress, and gap analysis to map how a position or thesis behaves across paths, not just the expected one.
Event and Macro Risk
Maintains earnings and macro calendars, scores event risk against historical distributions, and overlays scheduled catalysts onto open research.
Signal Validation
Gates every engine output through agreement checks, historical review, and go/no-go logic before it reaches the desk as a research candidate.
AI-Assisted Synthesis
Language models narrate and cross-examine deterministic engine output. They explain the research. They do not generate the numbers.
Multi-engine research architecture.
Data flows through shared ingestion and normalization, then fans out to independent engines. Their outputs are compared, challenged, and synthesized before a human makes the call.
Each engine is designed to answer a narrower question. NRGX becomes valuable when those answers are compared, challenged, and synthesized.
Simplified public view. Engine internals, scoring methods, and thresholds stay behind the door.
Built as infrastructure, not a collection of indicators.
NRGX Labs is a production software system before it is a research product. Every figure below has a concrete source in the codebase.
Measured directly from this repository · July 2026 · no trading or performance figures appear on this page.
Modular service architecture. Each engine lives in its own module with isolated scoring, screening, and review logic, registered through 22 routed service modules on a shared FastAPI core.
Shared schemas and normalization. Every engine reads the same validated, session-aware view of prices, surfaces, calendars, and events.
Six market-data integrations. Independent feeds across options analytics, equities, fundamentals, news, and macroeconomic series, so no single vendor defines the view.
Simulation and backtesting. Monte Carlo path simulation, historical replay, and stress harnesses sit beside the live engines, not in a separate research notebook.
Caching and state. Redis carries session state, engine caches, and market snapshots so research stays fast without re-fetching the world.
Scheduled research jobs. Cron-driven pipelines build calendars, pre-open snapshots, and weekly reviews before the desk sits down.
Tested continuously. More than 1,400 automated tests across 125+ test files cover engines, gating, schemas, and simulators.
AI kept in its lane. OpenAI models narrate and cross-examine deterministic output through a separate front layer. The numbers never come from the language model.
Selected research domains.
Each domain is framed as a research question. None of them promises an outcome.
RD/01 · VOL_SURFACE_ENGINEVolatility Surface
Examines how implied volatility is distributed across strikes and expirations to identify changes in risk pricing, convexity, and market expectations.
RD/02 · DEALER_GAMMA_CONTEXTGamma and Dealer Exposure
Estimates where dealer hedging pressure sits relative to spot, and how that positioning is likely to amplify or absorb directional moves.
RD/03 · REGIME_MODELMarket Regime
Classifies the prevailing volatility and trend environment, and asks the harder question: which relationships stop working when it changes.
Options Flow and Open Interest. Tracks open-interest clusters and flow concentration to locate the strikes where positioning, not narrative, is defining the field.
Skew and Term Structure. Watches how the market prices downside protection and time, and flags when those relationships depart from their own history.
Event and Macro Risk. Scores scheduled catalysts, earnings, and macro releases against historical outcome distributions rather than headline sentiment.
Earnings Volatility. Studies how implied volatility builds and resolves around earnings, and when the market has over- or under-priced the event.
Scenario Analysis. Simulates thousands of paths across price, volatility, and time to map how a structure behaves away from the expected case.
Signal Agreement and Conflict. Compares independent engine outputs on the same underlying question, treating disagreement as evidence rather than noise.
Research should reduce uncertainty, not manufacture confidence.
The system is designed to surface evidence, conflict, risk, and conditional opportunity. Human judgment remains central.
- No guaranteed outcomes, and no pretense of them.
- No single magic score. Every engine answer carries its context.
- No black-box claims of certainty. Inputs and logic stay inspectable.
- Signals remain conditional on regime, liquidity, and the calendar.
- Models are most useful when they disagree, so disagreement is preserved.
- Research that fails historical review does not reach the desk.
Built by Joshua b Smith.
A founder who has spent his career turning complex systems into operating companies, and then applied that discipline to markets.
Joshua b Smith is a six-time founder and operator whose work has crossed ecommerce, consumer products, supplements, manufacturing, subscription commerce, artificial intelligence, and financial research.
After more than 25 years building and operating companies, he began developing NRGX Labs as a private research environment for examining markets through software, data, and multiple independent analytical models. He conceived the system, directed the architecture, and built it.
The platform reflects the same operating philosophy behind his other systems: build the infrastructure, connect the data, expose the blind spots, and make better decisions.
Meet Joshua b SmithOne system inside a broader body of work.
NRGX Labs is one expression of a broader body of founder-built software spanning ecommerce intelligence, healthcare, Scripture, and quantitative market research.
RavenOS GROWTH
AI-assisted growth intelligence for ecommerce operators, built to surface the revenue leaks dashboards miss.
ravenos.io →InjuryOS LEGAL
AI-native signed-case acquisition infrastructure for plaintiff injury firms, from first touch to signed case.
injuryos.io →Versefold SCRIPTURE
A quiet Scripture-first Bible app for iPhone, designed around less phone and more Word.
versefold.app →Joshua b Smith FOUNDER
The founder arc behind the ecosystem: 25+ years of companies, exits, systems, and operating judgment.
joshuabsmith.io →Build enough independent intelligence to see what one model misses.
NRGX Labs is an active research and engineering project. The system continues to expand as new engines, datasets, validation methods, and analytical workflows are developed.