Co-Founder @NextHire-AI

Ajayaditya Lokchandra  • Peerlist

<p>Software development has changed a lot with AI and modern workflows, but hiring still relies heavily on DSA rounds and interview-style evaluations. We built NextHire to help companies assess developers through real engineering tasks, so they get a clear picture of how candidates actually think, work, and solve problems on the job.</p>

NextHire AI

Designed and developed a sleek, modern personal portfolio using Framer, focused on smooth animations, responsive design, and an intuitive user experience. Showcases creative work with a minimal, high-impact aesthetic.

Ajayaditya - Portfolio V2

Cocoon is a comprehensive home service booking platform designed to connect professionals, customers, and administrators through a seamless ecosystem. Built during IITM's online BS in Data Science coursework, the platform caters to three distinct user roles: service providers, customers, and administrators, ensuring smooth interactions and efficient service management. Key features include secure authentication using Flask Security, scheduled task handling with Celery and Cron, and real-time email notifications via Mailgun. The platform leverages Flask for the backend, SQLite3 for the database, Redis for caching, and Vue.js for an intuitive, responsive frontend. Awarded Capstone Project Honour, Cocoon stands out for its robust architecture and user-centric design, addressing real-world challenges in service booking and administration. While it’s not live, Cocoon remains a noteworthy demonstration of technical skill and project execution.

Cocoon - Discover Home Comfort

Your Second Brain, Reimagined Neuronote flips traditional note-taking on its head. Designed for thinkers, not checklist chasers — it takes raw, messy thoughts and turns them into structured, atomic notes with zero manual effort. Built-in semantic search, auto-tagging, and smart summaries mean you spend less time organizing and more time using your brain. No bloated dashboards. No “productivity rituals.” Just brutal efficiency with AI under the hood.

Neuronote

I showcase my interests, experience, skills, and projects all in one place. Built with React and Tailwind CSS, and hosted on Vercel, the site is currently optimized for desktop and responsive across different screen sizes. There are still some bugs on mobile that I’m working on, but improvements are on the way! Check it out here: https://suyesh-portfolio.vercel.app/

Personal Portfolio

Built a visual workflow builder for PMs to automate GitHub → Slack/Asana flows via drag-and-drop interface and templated triggers. Integrated Gemini API (1.5 Flash) to autofill config fields from GitHub payloads, reducing setup time by ~50%. Delivered MVP in 7 days; rebuilt V2 with responsive layout, redesigned UI architecture, and advanced AI tooling. Achieved 98% webhook delivery reliability and 95% performance score on Google Lighthouse audits

Quirk – GitHub Automation for Product Teams

Trade Reads is a book exchange platform that allows users to list books they own and trade them with others. Built with Next.js, TypeScript, and Drizzle ORM, the platform enables seamless book discovery and exchange through a user-friendly interface.

Trade Reads - Book Exchange Platform

- Architected and deployed a multi-service enterprise RAG platform integrating internal knowledge (Confluence) with validation guardrails, hybrid retrieval (Pinecone dense search + MS-MARCO MiniLM reranking), and intent-aware routing to deliver reliable, citable AI responses while reducing unnecessary vector calls and latency. - Designed a decoupled ingestion pipeline (webhooks → staged state → deferred re-indexing) with retry logic and idempotent upserts to enable non-blocking, scalable indexing across enterprise data sources. Implemented enterprise-grade trust and reliability layers using DeBERTa-MNLI–based contradiction detection, Guardrails output validation, and modular Dockerized services (LLM engine, indexer, reranker, NLI, analytics) with fail-safe logging and non-blocking error handling. - Instrumented full-pipeline observability using ClickHouse + Metabase to track stage-level latency, status, and trace IDs, and integrated PromptLayer + LangSmith for prompt tracing, evaluation, and performance monitoring. Demo Link - https://drive.google.com/file/d/1owk99_p45ifnRXTgqgfJuUVEkU6i58fo/view

Nucleus AI — Enterprise Knowledge RAG Platfor

Neuronote is an AI-powered knowledge structuring engine that converts unstructured text into typed entities (tasks, insights, people, temporal references) for structured retrieval and downstream reasoning. I designed a hybrid retrieval system combining semantic embeddings with lexical boosting and weighted scoring, later replacing external vector infrastructure with an in-app similarity pipeline to reduce infrastructure cost by ~50% while maintaining accuracy. The system includes dynamic similarity thresholding and reranking to balance recall against hallucination risk, along with a context-aware summarization pipeline via Gemini. Client–server synchronization was optimized through local caching and mutation reconciliation, reducing server load by ~40% while preserving data consistency.

Neuronote | Your Second Brain, Reimagined

Nucleus AI is a production-oriented RAG system designed for enterprise knowledge retrieval and reliable LLM responses. I architected a multi-service pipeline handling ingestion, embedding, hybrid retrieval, reranking, generation, and validation, with containerized services and clear system boundaries. The ingestion layer supports staged indexing, webhook-driven updates, retry logic, and idempotent upserts to prevent index corruption under iterative knowledge changes. Retrieval combines vector search with cross-encoder reranking and intent-aware filtering to improve relevance while reducing redundant inference calls. The system includes NLI-based validation and stage-level latency tracing (~7s local end-to-end) using structured logging to enable bottleneck isolation and reliability-focused optimization.

Nucleus AI | Enterprise Knowledge RAG Platfor

Quirk is an event-driven workflow automation engine that converts GitHub webhooks into structured execution across Slack and Asana via a graph-based runtime and PostgreSQL-backed state layer. I engineered idempotent webhook ingestion with retry semantics and asynchronous processing, achieving ~98% reliable execution under concurrent events. The system includes conditional routing and optional LLM-assisted payload mapping, reducing manual configuration time by ~50% and reflecting a production-focused approach to automation architecture.