Hardik Umesh Choksi

The long version

The work behind Blue River Labs.

1. Chapter 01 · 2008 — 2009

   A patent written by a student who refused to pay a bribe

   Hardik's journey began not with a market trend, but with a problem he experienced personally as an engineering student in India. Obtaining a household cooking-gas connection — an essential service supported by government subsidy — was unnecessarily difficult, and even after securing it, receiving the monthly cylinder often involved additional payments and dependence on individuals exercising unchecked discretion. What should have been a rightful household entitlement was frequently treated as a privilege.

   Studying the problem more deeply, he recognized a much larger structural failure: paper-based identity records, fragmented distributor systems, manually entered deliveries, weak inventory visibility, duplicate household accounts, unauthorized diversion of subsidized cylinders, and a parallel economy that benefited from opacity at multiple levels of the supply chain. His foundational insight was that the problem could not be solved by replacing one official, distributor, or delivery person with another. It required a neutral technological system that could not be bribed, threatened, influenced, or pressured to retaliate against a consumer.

   In December 2008, without the financial resources to retain a patent attorney, he prepared and filed his own patent application for a technology-driven system connecting consumer identity, entitlement, transaction verification, and cylinder-gas distribution. Patent: IN 2416/MUM/2008. Errors in the self-prepared filing ultimately delayed and stalled the patent approval, but the application became the conceptual blueprint for his first company.

   In February 2009, while still in the third year of his Electronics and Telecommunications Engineering program and barely twenty years old, he founded Systemkraft as its sole founder. Building it required recruiting engineers and specialists considerably older and more experienced than he was, at a time when a student founder leading senior professionals was far less culturally accepted in India. He built that early team around a principle that has continued throughout his career: hire people who know more than you in their specialties, learn from them, and unite their expertise around a systems-level mission.

2. Chapter 02 · 2009 — 2019

   Systemkraft: fuel distribution, national identity, and banking access

   Systemkraft turned the original idea into an integrated public-infrastructure platform connecting four essential pillars: government policy, public-sector enterprises delivering life-essential products, India's emerging biometric identity infrastructure, and the formal banking system. Its central product was an end-to-end LPG Distribution Management System developed for environments associated with Bharat Petroleum Corporation Limited, Hindustan Petroleum Corporation Limited, and Indian Oil Corporation Limited — Government of India enterprises collectively serving more than 150 million household LPG accounts through thousands of distributors, with each account generally representing an entire family.

   Instead of treating customer onboarding, identity, inventory, delivery, complaints, banking, and subsidy entitlement as unrelated applications, Hardik designed them as one secure and auditable operating architecture. Systemkraft developed cloud-based customer, booking, inventory, inspection, complaint, and service modules; local distributor-management software; IVRS and customer kiosks; RFID and magnetic-stripe smart cards; handheld point-of-sale terminals capable of operating online or offline; GPS- and GPRS-enabled delivery verification; biometric-device integrations; electronic KYC; duplicate and exception management; low-connectivity transaction storage and synchronization; banking-onboarding workflows; and data exchange with SAP and other central oil-company systems.

   In September 2010, Systemkraft entered into an agreement with the Unique Identification Authority of India covering Aadhaar enrollment, project-specific electronic KYC, authentication, and support for the BPCL, HPCL, and IOCL distribution-management project. This was not an integration with a mature, ready-made eKYC service: between approximately 2010 and 2013, Systemkraft was helping build and field-test an early electronic identity-verification layer while India's national authentication ecosystem itself was still developing. The result was a closed operational chain connecting a real person, a legitimate household account, a distributor, inventory and cylinder movement, booking and delivery confirmation, a service or complaint record, and, where required, the bank account through which a government benefit could be received.

   The scale reflected in the record is extraordinary. A BPCL certificate recognizes a national-scale system deployed across more than 6,000 distributors, encompassing 18 mission-critical operating modules, running with near-zero downtime, and affecting nearly 50 million customers. The HPCL project record confirms Aadhaar-linked onboarding for more than 30 million customers and estimated prevention of nearly INR 150 crore in annual diversion or leakage within that network. The IOCL record identifies service support for approximately 40 million customers.

   Taken together, the three oil marketing companies represented a combined LPG ecosystem of more than 150 million household accounts and beyond 20,000 distributors — enterprises whose FY 2025-26 revenues total approximately US$199.5 billion. Based on the implementation record reviewed in the BPCL recommendation letter, these workflows helped enable first-time formal bank-account access for more than 50 million people in rural India. They became part of a broader national transformation that eliminated approximately 40.9 million duplicate, fake, nonexistent, or inactive LPG connections and generated more than INR 74,031 crore in cumulative estimated benefits through March 2025. Systemkraft's role was the early last-mile technology and operating layer that made identity-linked customer cleansing, electronic KYC, verified delivery, bank-account seeding, and transparent subsidy administration workable in actual field conditions.

3. Chapter 03 · 2015 — 2020

   SSRI Creative Construction: graph intelligence for asset-intensive infrastructure

   Building Systemkraft alone shaped how Hardik approached his second company. Having understood the burdens and blind spots of solo entrepreneurship, he co-founded SSRI Creative Construction with a deeply experienced infrastructure professional whose civil, structural, metro, tunneling, and power-sector knowledge complemented his own software architecture, data modeling, machine-learning, and product background. SSRI was not a conventional construction contractor; it was an infrastructure-technology company created to convert the operational complexity of metro railways, power-generation facilities, underground construction, industrial projects, and oil-and-gas construction environments into a connected enterprise operating system.

   Hardik led the progression from SSRI ERP 1.0 to SSRI ERP 2.0, building a graph-based platform in which projects, sites, assets, components, work orders, maintenance schedules, vendors, spare parts, permits, inspections, approvals, safety events, failures, costs, and field evidence were modeled as connected entities rather than isolated database records. Before large language models became mainstream, he incorporated machine learning and Graph Neural Networks to evaluate the operational neighborhood surrounding an asset, update health and risk scores, rank work-order urgency, assess vendor and dependency risk, forecast material or spare-part constraints, and recommend practical actions.

   The platform was adopted by Reliance Infrastructure, Mumbai Metro One, Adani Electricity, and the Dahanu Thermal Power Station. The Mumbai Metro One deployment covered an approximately 11.4-kilometre corridor, 12 operating stations, more than 1,000 authorized users, and approximately 10,000 configured asset records. At Dahanu, the contracted architecture was designed to govern as many as 15,000 assets across two generating units and their associated boiler, turbine, generator, coal-handling, ash-handling, switchyard, water-treatment, fire, workshop, and stores systems.

   Reliance Infrastructure and Mumbai Metro One conferred the 2019-2020 Revolutionary Technology of the Decade award on SSRI. The citation described SSRI ERP 2.0 as a first-of-its-kind operating system platform for heavy infrastructure and expressly recognized its use of machine learning, graph neural networks, and graph-based technology, reporting nearly 14 hours of recurring manual work eliminated each week, more than 800 team-hours saved per team annually, incident-prediction accuracy in the 80% to 90% range, and complete accuracy in defined inventory-forecasting scenarios.

4. Chapter 04 · 2019 — 2022

   The Choksi Group: turning a distribution monopoly into a performance system

   SSRI was subsequently acquired by Surat Gas Transport Company, a Choksi Group company. Following the acquisition, Hardik joined the group — whose owners are unrelated to him — as Vice President of Strategy and Operations and Head of Engineering. He entered a conglomerate described in the recommendation record as generating approximately US$3.4 billion in annual revenue, employing more than 11,500 people, and operating an LPG network of more than 900 distributors across 22 bottling plants with an aggregate capacity of roughly two million metric tons per year.

   He re-architected that network from an exclusive geographic model, in which one distributor controlled most functions within a territory, into a performance-based functional network organized around sales, customer service, and warehousing and logistics. Multiple qualified participants could compete within the same area and function, with business allocated according to response time, delivery completion, inventory availability, complaint closure, safety performance, and service reliability. He then built the graph-intelligence stack connecting customers, warehouses, distributors, vehicles, inventory, delivery routes, personnel, complaints, safety events, and completed deliveries.

   Within that network, average complaint-resolution time fell from 11 hours to 1.8 hours and average fulfilment time — from a customer requesting supply to completed delivery — fell from 32.5 hours to 5.1 hours, both reductions of approximately 84%. Gas-leakage incidents and related complaints declined by roughly 50%. He had also anticipated the resistance such a reform would create: nearly 150 distributors chose to leave rather than meet the new standards, but more than 200 had already been brought in, leaving the network with a net gain of over 50 distributors through the transition.

5. Chapter 05 · 2022 — 2026

   Building the intelligence layer

   After repeatedly building consequential systems across identity, financial inclusion, infrastructure, energy, and national supply chains, Hardik made a deliberate decision to return to university before building his next global company. The objective was not another credential. He wanted to immerse himself in the world's most concentrated innovation ecosystem, develop a lifelong network of founders, researchers, investors, and operators, and reconsider from first principles what could be built for enterprises worldwide. After receiving admission opportunities from Harvard, MIT, and Stanford, he chose Stanford University, where he completed a Master of Science in Management in 2023.

   At Stanford, he extended his engineering and graph-intelligence background into unconventional applications of artificial intelligence, particularly agriculture and the fresh-food supply chain. His research examined how farmers, suppliers, products, transportation, stores, restaurants, recipes, inventory states, demand, shelf life, substitutions, pricing, waste, and physical outcomes could be represented as a connected, continuously learning operating system. That research became the foundation for Royal Fresh, an applied intelligence platform intended to make perishable commerce more predictable, fresher, less wasteful, and more economically sustainable.

   He also explored the implementation of artificial intelligence across complex and regulated sectors, including manufacturing, healthcare, banking, insurance, government, defence, energy, agriculture, fresh food, hospitality, climate, and software engineering — and recognized that although these industries differ in terminology, systems, regulations, and workflows, their foundational intelligence challenge is remarkably similar. Organizations must connect fragmented context, resolve real-world entities, understand relationships over time, predict outcomes, reason within constraints, govern actions, and learn from human decisions.

6. Chapter 06 · 2026 — today

   Blue River Labs: generalizing the architecture

   Today, as Founder and Chief Executive Officer of Blue River Labs Inc., Hardik is generalizing the architectural principles that have defined his career. Blue River Labs is building a horizontal enterprise intelligence layer that connects fragmented code, documents, data, communications, incidents, telemetry, operational systems, decisions, and human knowledge; resolves them into a canonical and temporal enterprise knowledge graph; predicts risks, likely outcomes, and propagation paths through its F3M graph model; transforms structured intelligence into reasoning context through its Graph Reasoning Interface; supports grounded reasoning through F2RM; and governs execution through its Agent OS and F2 Cockpit.

   Rather than repeatedly sending large volumes of fragmented context through isolated model calls, the platform is designed to be token-efficient and reusable across domains. Its initial proving grounds are an AI-native software development lifecycle and Royal Fresh.

   The continuity across Hardik's journey is direct: Systemkraft connected identity, entitlement, physical delivery, and banking; SSRI connected assets, workflows, dependencies, and risk; The Choksi Group connected an entire physical distribution network; and Blue River Labs is now building the intelligence foundation through which enterprises can understand relationships, predict consequences, reason over constraints, act with permission, and remain accountable.