Silver Touch Technologies Wins Contract from RITES to Build AI-Native DPR Appraisal Platform

Silver Touch Technologies Wins Contract from RITES to Build AI-Native DPR Appraisal Platform​

Silver Touch Technologies Limited has secured a major order from RITES Limited, a Navratna Government Enterprise under the Ministry of Railways, for the Design, Development, and Implementation of PARAKH. This project involves creating India's first self-hosted, multi-model AI platform dedicated to infrastructure project appraisal, positioning Silver Touch as a key player in national infrastructure governance using advanced artificial intelligence.

The contract with RITES Limited aims to build a specialized AI-Based Detailed Project Report (DPR) Appraisal and Intelligence Platform known as PARAKH. This platform is designed not merely as an application with AI features, but as a ground up AI native system intended to transform how India's largest infrastructure consultant assesses DPRs.

The engagement involves architecting and deploying the sophisticated AI platform for RITES Limited. The project timeline is set at 24 months, and the valuation of this order amounts to Rs. 6,27,76,000.

Key details of the contract awarded by RITES Limited are provided in the table below:

Building an AI Native Infrastructure Intelligence Tool​

The scope of the work under this engagement is extensive, requiring Silver Touch to deliver a production grade solution that incorporates cutting edge technology. The PARAKH platform integrates several specialized components:

• Large Language Models: It utilizes three self-hosted open source LLMs (Llama 3.1 70B, Mistral 7B, and Qwen 2.5 7B) deployed within RITES’ secure data center with zero external data egress.
• Validation and Intelligence: The system includes over 500 codified Engineering Validation Rules derived from standards such as IRC, RDSO, IS, and MORTH.
• Knowledge Retrieval: It features a Hybrid Retrieval Augmented Generation (RAG) framework that combines semantic vector search with keyword search across more than 1,000 historical DPRs to ensure grounded, citation backed appraisal.
• Security and Quality: A Four Layer Hallucination Prevention Framework mandates that every AI generated insight is source verified, confidence scored, citation mandatory, and human approved.

The platform is also designed to handle massive data inputs, processing DPRs ranging from 500 to 5,000 pages using Document Intelligence Pipeline capabilities involving OCR, table extraction, and engineering entity recognition. Furthermore, it integrates GIS with the PM GatiShakti National Master Plan and includes BIM/IFC interoperability.

Significance and Domain Moat​

Silver Touch Technologies emphasizes that this project represents a strategic shift toward becoming an AI native platform company rather than solely an IT services provider. PARAKH is described as a mission critical, production grade system where artificial intelligence is the central value proposition, addressing the infrastructure sector's immense needs.

This contract grants Silver Touch a significant domain moat in the AI for Infrastructure space. The proprietary nature of the 500+ codified engineering rules and the historical DPR knowledge base provides a compounding competitive advantage that increases with every appraisal conducted by the platform.

Minesh V Doshi, Executive Director at Silver Touch Technologies Limited, stated that he is honored to partner with RITES Limited on building what they view as India’s most advanced AI native infrastructure intelligence platform. He noted that the deployment of self-hosted LLMs and a 500 rule engineering validation engine within RITES' secure infrastructure establishes Silver Touch as a pioneer in AI for infrastructure governance.

Silver Touch Technologies Limited is an established IT and digital transformation company with operations across India, the UK, the US, and Canada, specializing in eGovernance, Cloud, SAP, AI/ML, and Cybersecurity solutions.