Hey, I'm
Mayank Bohra
Co-Founder, Highlyt · GenAI Engineer. Building Highlyt, a PDF reader that plugs into Claude and ChatGPT via MCP, while continuing as Founding Engineer at Gradeless AI on Rehearsal AI. Writing about the messy reality of shipping GenAI.
I'm Mayank Bohra, Co-Founder of Highlyt and Founding Engineer at Gradeless AI. At Highlyt, we're building a typed, cross-document knowledge graph for readers who live inside Claude and ChatGPT. At Gradeless, I work on Rehearsal AI, a voice-powered AI interview prep platform serving 5,000+ candidates.
My Story
I'm Mayank Bohra, Co-Founder of Highlyt and Founding Engineer at Gradeless AI. At Highlyt, we're building a PDF reader and typed knowledge graph that plugs into Claude and ChatGPT via MCP, so AI can read from and write to your highlights directly. I handle the technical side; my co-founder Rajika Patel leads design and marketing.
At Gradeless AI, I work on Rehearsal AI, a voice-powered AI interview prep platform serving 5,000+ candidates across IIMs, engineering placements, and campus recruitment. I architected the interview backend and built DeepProbe, the adaptive questioning engine.
My focus is consistent across both active roles: turning messy, open-ended problems into production LLM systems that feel simple for non-technical users. At Highlyt the surface is reading and annotation; at Rehearsal it is voice and interviewing. The engineering discipline is the same.
Earlier, I built AI tools at Jaipuria Institute of Management, India's first AI-Native B-School, including interview platforms serving 2,000+ students across 4 campuses and natural-language interfaces for querying academic content.
I write about what I learn while building: LLM engineering, MCP servers, Claude workflows, and the messy reality of shipping GenAI products. My writing appears on Towards AI, FAUN.dev, and my Substack.
What I'm Focused On
Work Experience
December 2025 – PresentCurrent
Co-Founder
Highlyt, highlyt.app · Co-Founder
Co-Founded Highlyt with Rajika Patel. Highlyt is a PDF reader and typed, cross-document knowledge graph that plugs into Claude and ChatGPT via MCP. I lead engineering, architecting the five-client platform (web, Chrome extension, mobile, MCP OAuth server, MCP API-key server) that share one FastAPI + Supabase backend. Built the knowledge-graph layer, typed highlight links (supports, contradicts, expands, questions, related), the Reading Assessment engine, and 27 MCP tools that let LLMs operate inside a reader's library. Rajika handles design and marketing.
November 2025 – PresentCurrent
Founding Generative AI Engineer
Gradeless AI · Full-time
Founding engineer building Rehearsal AI, a voice-powered interview preparation platform serving 5,000+ candidates across IIMs, engineering placements, and campus recruitment. Architected the interview backend with minimal latency voice interactions and built DeepProbe, an adaptive questioning engine that prevents repetition through domain-aware rotation while forcing deeper responses. Designed the question delivery system with controlled randomization across different areas specific to different domains, real-time answer analysis, and credibility scoring.
December 2024 – Present
AI Developer
Jaipuria AI Labs, Jaipuria Institute of Management · Full-time
Developed AI-powered educational tools at India's first AI-Native B-School. Created interview simulation platform serving 3,000+ students across 4 campuses. Created natural-language interfaces over Supabase/PostgreSQL for querying academic content.
August 2023 – December 2024
Teaching Assistant
JK Lakshmipat University · Part-time
Provided hands-on assistance to students during lab sessions. Assisted in the evaluation of students' performance.
May 2023 – December 2024
Technical Content Writer
GeeksforGeeks · Freelance
Wrote technical articles on programming concepts, data structures, and algorithms. Contributed to educational content reaching millions of developers and students worldwide.
June 2023 – July 2023
Junior Software Engineer
Dexpert Systems PVT. LTD. · Internship
Built a Progressive Web App using ReactJS, developed a Management System from scratch with ASP.NET and MS SQL Server, upgraded a Flutter mobile app, and created dynamic dashboards for payables/receivables using Java Struts 2.0 and MySQL.
Academic Background
B.Tech in Computer Science Engineering
JK Lakshmipat University, Jaipur
Case Studies
Not just what I built, but how I think. These case studies show my engineering decisions, trade-offs, and learnings.
Building a PDF Reader for the LLM Era
Highlyt · highlyt.app
Why I don't want an AI that reads for me, I want one that reads with me. A case study in building a typed, cross-document knowledge graph that Claude and ChatGPT can read from and write to via MCP.
Scaling AI Interview Simulations
Rehearsal AI @ Gradeless
How I architected a voice-powered interview platform serving 5,000+ candidates across IIMs, FAANG prep, and campus placements, from problem space to production.
Projects
Side projects, experiments, and open source contributions.
How I Think
The principles and trade-offs that guide my engineering decisions.
Ship Fast, Learn Faster
I'd rather ship something imperfect and iterate than build the perfect solution no one uses. Real user feedback beats theoretical optimization every time.
Trade-offs Over Best Practices
There's no universal "best" solution. I focus on understanding constraints, evaluating trade-offs, and choosing what's right for this context, not what's trendy.
Document Decisions, Not Just Code
Code tells you what. Comments tell you how. Documentation tells you why. Future me (and teammates) need all three to maintain systems effectively.
Build for Non-Technical Users
The best AI systems disappear into the background. My goal is making complex technology feel simple and intuitive for the people who actually use it.
My Tech Stack
Backend
Frontend
AI / LLM
Tools
Testimonials
"I find Mayank a versatile talent with fantastic technical acumen and zeal to get the task done to satisfaction. Wishing him all the best for his future.
"Mayank is among the top students in the class. He has been a dedicated and hard-working student. He has often shown professional skills even as a student and has been goal-oriented. He doesn't need a lot of hand-holding. Even outside class, he is very supportive in nature. There to help his friends and peers. I wish him all the best in his career. I am sure he will become an asset to any company he joins.
"Mayank is a very good programmer. He has very good technical skills in Java and Data Structures.
Thoughts & Insights
I write about what I learn while building: LLM engineering, Claude workflows, and the messy reality of shipping GenAI products. Featured on Towards AI, FAUN.dev, and other publications.
Work With Me
I help startups and institutes build AI-powered products, from interview simulations to internal chatbots. If you need someone who can turn messy requirements into production systems, let's talk.
AI Chatbots & Assistants
Custom chatbots for institutes, SMEs, or internal tools. RAG-powered, context-aware, production-ready.
Interview Simulation Tools
Voice-powered interview prep systems for placement cells, coaching institutes, or HR teams.
FastAPI + LLM MVPs
Quick 2-3 week builds for founders who need an AI-powered MVP to validate their idea.
Follow My Writing
I share what I learn while building production AI systems: LLM engineering, voice AI, and the messy reality of shipping GenAI products.
Read on MediumLet's Connect
Building in GenAI, MCP, or reading tools? Want to chat about Highlyt, LLMs, or career stuff? I'd love to hear from you.
