# triNetra Research triNetra Research is an independent research organisation studying the structural layer of consequential AI decision-making. Research question: when an AI system makes or informs a consequential decision, can the reasoning behind that decision be preserved, verified, and explained? Based in New Delhi, India. Founded 2026. Research Programme: EAD Research Programme (four published working papers: EAD-2026-01, EAD-2026-02, EAD-2026-03, EAD-2026-04). All four papers are available on SSRN. ## Researcher Shubham Agarwal Founder, triNetra Research New Delhi, India ORCID: https://orcid.org/0009-0003-0822-4818 SSRN: https://ssrn.com/author=12335668 LinkedIn: https://www.linkedin.com/in/shubham-agarwal-trinetra/ Email: research@trinetra.life Profile: https://trinetra.life/researcher ## Eagle Framework The Eagle Framework is triNetra's deterministic structural assessment framework: 35 patterns across 7 auditability dimensions. It makes the reasoning behind consequential decisions traceable by turning institutional structure into an explicit record. Derived from the EAD Research Programme via the PaaF methodology. Seven dimensions: Evidence Attribution, Decision Traceability, Confidence Calibration, Counterfactual Accountability, Human Oversight Readiness, Incident Reconstruction, Audit Trail Completeness. Full reference: https://trinetra.life/frameworks ## Methodology: PaaF (Problem and Framework Methodology) PaaF is triNetra's recursive structural research methodology. It operates through cycles of observation, pattern extraction, structural interpretation, and evolution direction. Each cycle either resolves the current constraint or reveals a deeper one. The process continues until the system's governing structure becomes visible. The methodology does not assume that the first identified problem is the actual problem. Four phases: 01 Problem Identification Research: Research begins at the symptom layer and works downward through process, structure, and dependency until the source condition is located. Not what is failing. Why it was always going to fail. 02 Pattern and Framework Design: Patterns are derived from observed system reality, not theory or best practice. Recurring structures, dependency chains, and failure modes are extracted and encoded into evaluation design. 03 Structural Interpretation: The framework is interpreted for the teams who will govern and evolve it. Contradictions, survivability pressures, and trajectory direction are made explicit. 04 Evolution Direction: Structural gaps are identified before they become failures. Adaptable pathways are designed for the framework to grow without losing structural coherence. ## Research ### EAD Research Programme Four published working papers on AI economics, organisational governance, and field research. All available on SSRN. ### EAD-2026-01: External AI Dependence and Startup Financial Survivability Author: Shubham Agarwal (ORCID: https://orcid.org/0009-0003-0822-4818) Research question: how does dependence on externally owned AI infrastructure affect the financial survivability of independent businesses? Proposes the External AI Dependence (EAD) framework, the EAD Index (EADI), and the Specialised Digital Asset hierarchy. Status: Working Paper. July 2026. SSRN: https://ssrn.com/abstract=7104058 DOI: https://doi.org/10.XXXXX/PLACEHOLDER URL: https://trinetra.life/ead-report ### EAD-2026-02: The Judgment Layer. Inductive Theory of Understanding Synthesis Failure Author: Shubham Agarwal (ORCID: https://orcid.org/0009-0003-0822-4818) Research question: why do large enterprises fail recurrently in domains where they possess relevant operational systems, qualified personnel, and regulatory compliance? Proposes the Judgment Layer as the missing synthesis layer between operational intelligence and executive judgment. Grounded in 15 empirical cases, USD 80B+ documented consequences. Status: Working Paper. July 2026. SSRN: https://ssrn.com/abstract=7103978 DOI: https://doi.org/10.XXXXX/PLACEHOLDER URL: https://trinetra.life/judgment-layer-theory ### EAD-2026-03: The Infrastructure Loop Author: Shubham Agarwal (ORCID: https://orcid.org/0009-0003-0822-4818) Research question: if Specialised Digital Asset revenue is directed back into infrastructure ownership and community knowledge transfer, what changes in the structure of the AI economy? Proposes the Infrastructure Loop mechanism across 11 assessed economies. Status: Working Paper. July 2026. SSRN: https://ssrn.com/abstract=7104079 DOI: https://doi.org/10.XXXXX/PLACEHOLDER URL: https://trinetra.life/infrastructure-loop ### EAD-2026-04: PaaF in the Field. A Structural Reading of the Current AI Builder Landscape Author: Shubham Agarwal (ORCID: https://orcid.org/0009-0003-0822-4818) Research question: when PaaF methodology is applied to direct field observation of independent AI builder communities, what structural conditions become visible? 18 observations across 4 analytical layers (Surface, Cognitive, Systemic, Developmental). Single structural finding: judgment layer absent at every level. Field evidence base for EAD-2026-01, EAD-2026-02, EAD-2026-03. WIPO WIIH 2026: USD 10 trillion global intangible investment. Status: Working Paper. July 2026. SSRN: https://ssrn.com/abstract=7104138 DOI: https://doi.org/10.XXXXX/PLACEHOLDER URL: https://trinetra.life/paaf-field-analysis Full text (Markdown): https://trinetra.life/PaaF_in_the_Field_EAD-2026-04.md ### JLT Adversarial Panel Review Complete pre-submission adversarial review of Judgment Layer Theory by six independent senior scholars. Three fatal weaknesses identified and resolved. Editorial decision: major revision with resubmission invitation. Status: Published July 2026. URL: https://trinetra.life/judgment-layer-adversarial ## Open Engineering The public implementation layer of the triNetra research ecosystem, housed at github.com/RocKing000/Portfolio. GitHub: https://github.com/RocKing000/Portfolio URL: https://trinetra.life/open-engineering The repository demonstrates that the triNetra lifecycle is complete and verifiable, from research to theory to architecture to engineering to reference implementation. ### Reference Implementations (5 systems) DIRE-X: Geopolitical intelligence platform. React, Vite, Node.js, Python, FastAPI, Supabase, Leaflet, Chart.js, Recharts. KYC Classification: Computer vision identity verification pipeline. Python, OpenCV, scikit-learn, MediaPipe, FastAPI, TensorFlow, PostgreSQL. Research connection: EAD-2026-02 (high-consequence automated decisions). LOS (Loan Origination System): Enterprise credit decisioning. React, .NET 8, C#, Docker, gRPC, RabbitMQ, Redis, PostgreSQL, SQL Server. Research connection: EAD-2026-02 (consequential financial decisions). SDLC Workflow: AI-assisted software delivery platform. React, TypeScript, Node.js, LangGraph, LangChain, MongoDB, Docker, Redis. Research connection: EAD-2026-04 (AI builder landscape structural reading). Solution Architecture: Multi-tenant enterprise analytics platform. Angular, TypeScript, .NET 8, Azure Blob Storage, PostgreSQL, Redis, SignalR. Research connection: EAD-2026-02 (enterprise decision support / Judgment Layer domain). ### Research Documentation 65+ architecture, methodology, platform, and case study documents in the triNetra/ documentation layer within the repository. Includes architecture specs, implementation contracts, evaluation documents, market validation, and technical references. ### Authored Works Three authored works included: RiskOpsBench 100 Sample Corpus, Illustrative Eagle Assessment, Manifest Evolution Report. ## Key Pages - [Homepage](https://trinetra.life): triNetra Research. Independent research organisation studying the structural layer of consequential AI decision-making. Research Programme: EAD Working Papers. - [Researcher Profile](https://trinetra.life/researcher): Shubham Agarwal. Founder and researcher, triNetra Research. ORCID, publications, research interests, biography. - [Research Library](https://trinetra.life/research): EAD Research Programme and independent studies. Four published working papers. - [EAD Research Programme](https://trinetra.life/ead-programme): Programme landing page. Four papers: EAD-2026-01, EAD-2026-02, EAD-2026-03, EAD-2026-04. - [Judgment Layer Theory](https://trinetra.life/judgment-layer-theory): EAD-2026-02. Inductive theory of understanding synthesis failure. 15 empirical cases, USD 80B+ consequences. Working paper. - [The Infrastructure Loop](https://trinetra.life/infrastructure-loop): EAD-2026-03. Self-reinforcing mechanism connecting Specialised Digital Asset revenue to infrastructure ownership participation. Working paper. - [PaaF in the Field](https://trinetra.life/paaf-field-analysis): EAD-2026-04. PaaF methodology applied to 18 field observations of the AI builder landscape. Judgment layer absent at every level. Field evidence base for EAD programme. Working paper. - [External AI Dependence Working Paper](https://trinetra.life/ead-report): EAD-2026-01. How external AI dependence affects startup financial survivability. EAD framework, EADI, Specialised Digital Asset hierarchy. Working paper. - [Research Methodology](https://trinetra.life/paaf): PaaF (Problem and Framework Methodology). triNetra's recursive structural research methodology. - [Eagle Framework Hub](https://trinetra.life/frameworks): The Eagle Framework. 35 patterns, 7 auditability dimensions. Full reference. - [Applied Analysis: India AI Governance](https://trinetra.life/applied-india-ai-governance): Eagle Framework applied to India's national AI governance architecture. Seven structural dimensions. - [Open Engineering](https://trinetra.life/open-engineering): Public implementation layer of the triNetra research ecosystem. Five reference implementations, 65+ documentation files, authored works. github.com/RocKing000/Portfolio. - [Contact](https://trinetra.life/contact): Email research@trinetra.life or WhatsApp +91 95282 15988. New Delhi, India. ## Contact Founder: Shubham Agarwal ORCID: https://orcid.org/0009-0003-0822-4818 Email: research@trinetra.life WhatsApp: +91 95282 15988 Location: New Delhi, India Coordinates: 28.6139 N, 77.2090 E