A tech-savvy innovator who loves coding and bringing creative ideas to life, constantly building MVPs and startup

github

Pratyush Rastogi • Peerlist

<h3>About Pratyush Rastogi</h3>Pratyush Rastogi is a Full-Stack Developer, Software Engineer, and Frontend Engineer with a specialization in AI Integration and Prompt Engineering. He focuses on manipulating Large Language Models (LLMs) to shape high-quality digital experiences and builds intentional systems using AI-assisted workflows. His approach emphasizes engineering correctness and design taste, delivering production-ready solutions with technical depth.<h3>Philosophy</h3>Pratyush believes the core of development has shifted from simply typing code to judgment and system design. He designs systems, directs AI agents in their creation, and remains accountable for all production decisions. This approach, termed 'agentic engineering,' orchestrates LLMs across the full development loop—plan, build, verify—while maintaining human oversight on architecture, correctness, and aesthetic. He emphasizes building guardrails for speed and safety, including typed contracts, tests, review gates, and robust tooling between agents and production, ensuring outputs are intentional and human-crafted.<h3>Selected Work</h3><ul><li>Backstop: Creator &amp; Maintainer. An open-source safety gateway that classifies risky PostgreSQL queries via AST parsing, gates destructive writes, and restores table snapshots. (Go, Python, TypeScript, PostgreSQL, Docker)</li><li>PulseBoard: Full-Stack Engineer. Sentiment analysis for Slack conversations, surfacing team mood and engagement signals through a real-time analytics dashboard. (Next.js, Node.js, Express, OpenAI, Slack)</li><li>RSS Posting (rssposting.com): Full-Stack Engineer. Production platform for automated Facebook and Instagram management with RSS feed posting, admin approval, and JWT security. (Spring Boot, JWT, CI/CD, React)</li><li>Not Just ToDo AI: Personalised project idea generator using prompt engineering for students. (Next.js, Spring Boot, MySQL, OpenAI)</li><li>Automated EDA Platform: Data analysis platform reducing statistical reporting time by 70% with interactive visualizations and automated PDF generation. (Python, Streamlit, Pandas, EDA)</li></ul><h3>Competencies</h3><ul><li>Languages: Java, C#, Python, SQL, HTML/CSS, JavaScript, TypeScript</li><li>Technologies: Spring Boot, React, Next.js, GraphQL, MySQL, MongoDB, REST APIs, Streamlit</li><li>Tools &amp; DevOps: Git, Docker, AWS, Linux, n8n, CI/CD Pipelines, OAuth API</li><li>Core AI: Integration, Prompt Engineering</li></ul><h3>Journey</h3><ul><li>Freelance Automation Engineer (Jun 2025): Delivered end-to-end monitoring workflows and integrated real-time Slack alerting.</li><li>Freelance Full Stack Developer (Oct 2025 - Dec 2025): Implemented an automated social media management platform.</li><li>Software Engineer I, Wednesday Solutions (Feb 2026 - Apr 2026): Shipped production features for PulseBoard and a Queue Management System.</li></ul><h3>Education &amp; Certifications</h3><ul><li>B.Sc. in IT, Graphic Era University (2022 — 2025)</li><li>Certifications: Data Science, Oracle Java (In Progress), AWS Solutions Architect (Currently Learning)</li></ul>Contact Pratyush at rastogipratyush630@gmail.com. Open for collaboration &amp; roles.

Pratyush Rastogi | Software Engineer

founder of FinPulse & SegmentAI

SegmentAI is a cutting-edge platform designed to revolutionize how businesses interact with and leverage artificial intelligence. It offers a comprehensive suite of tools and services aimed at simplifying the integration and management of AI models into existing workflows. SegmentAI empowers users to harness the power of AI for enhanced decision-making, process automation, and customer engagement. The platform focuses on providing intuitive interfaces and robust backend infrastructure to support a wide range of AI applications, from data analysis and predictive modeling to natural language processing and computer vision. With SegmentAI, businesses can unlock new levels of efficiency and innovation by seamlessly incorporating advanced AI capabilities into their operations.<h3>Key Features:</h3><ul><li>AI Model Integration: Easily connect and deploy various AI models.</li><li>Data Analysis Tools: Gain deeper insights from your data with advanced analytics.</li><li>Process Automation: Automate repetitive tasks and streamline workflows.</li><li>Customer Engagement: Enhance customer interactions with AI-powered solutions.</li><li>Scalable Infrastructure: Built to handle growing AI demands.</li></ul>SegmentAI is committed to making AI accessible and actionable for businesses of all sizes, driving growth and competitive advantage in the digital age.

SegmentAI

<h3>Problem: The PocketOS Incident — April 25, 2026</h3>A founder was using Cursor + Claude to fix a credential issue in staging. He didn't ask the agent to touch production. He didn't ask it to delete anything.In 9 seconds, the agent wiped the entire production database — and the backups, which were stored on the same volume.The agent's own logs later admitted it knew the rule. It broke it anyway.Railway's CEO personally restored the data two days later. The company got lucky.The gap this exposed isn't unique to Railway.The same failure mode — an agent with credentials doing something irreversible without asking — happens to databases constantly, just with SQL instead of a curl command.<h3>Database Safety for Developers and AI Agents</h3>Backstop provides a critical layer of defense for your production databases, especially when working with AI agents. It acts as a PostgreSQL-aware gateway, intercepting and classifying risky SQL queries before they can cause damage. Key features include:<ul><li>Query Interception and Classification: Analyzes SQL queries in real-time using AST parsing to determine risk levels (SAFE, HIGH, CRITICAL).</li><li>Approval Workflows: Risky write operations can be configured to require human approval before execution.</li><li>Recovery Readiness Verification: Ensures that recovery points are verified before critical operations are allowed.</li><li>Auditable Restore Paths: Provides auditable mechanisms to restore table snapshots from your own storage.</li><li>Agent Identity Tracking: Assigns an actor identity to AI agents, enabling attribution and auditing of their actions.</li><li>Fast Table Restore: Offers a preview-first restore process, recovering data into a separate table within your infrastructure.</li><li>Bring Your Own Storage (BYOS): Supports snapshot storage on AWS S3 or S3-compatible endpoints like MinIO.</li><li>Open Source Core: The core components, including the SDK, gateway, and restore engine, are Apache-2.0 licensed and available for self-hosting.</li></ul>Backstop is designed to integrate seamlessly with your existing stack, supporting various databases, programming languages, and agent frameworks. It offers a self-hosted open-source core, with plans for commercial workflow layers and enterprise features.