About Skills Projects Experience Contact
Available for opportunities

Mayank
Bhadrasen

AI Automation Engineer · n8n Workflow Builder · Former Frontend Lead @VeraAI

Bridging the gap between long-horizon AI agents and production-ready user experiences. Specializing in MCTS, automated workflow pipelines, and interactive 3D frontend engineering.

View Work LinkedIn
scroll to explore
Background

Engineer at the
AI frontier

I'm Mayank Bhadrasen, an MS in Information Systems graduate from Northeastern University who went from SDE Intern to Frontend Lead at VeraAI Technologies by learning fast and shipping faster.

My work sits at a rare intersection: I understand the mathematics of agentic AI—Monte Carlo Tree Search, LLM benchmark harnesses, workflow orchestration—and I know how to surface that complexity as clean, intuitive interfaces.

Whether it's architecting a TextCraft benchmark system with full checkpointing, replicating SWE-SEARCH with a perfect test suite, or pioneering 3D visualization for construction workflows, I build things that actually work at scale.

12/12
SWE-SEARCH test suite pass rate
3D
Three.js visualization lead at VeraAI
MS
Information Systems, Northeastern University
n8n
Workflow automation & AI pipeline builder
Technical Arsenal

What I work with

From agentic systems to interactive frontends—my stack spans the full AI-to-UI pipeline.

🤖
AI & Automation
n8n GPT-5 APIs Agentic Systems LLM Integration AgentGym-RL AI Benchmarking Prompt Engineering
⚛️
Frontend & Visualization
React Three.js JavaScript TypeScript WebGL Tailwind CSS 3D Rendering
🔧
Backend & Systems
Python MCTS Git Automation Testing Frameworks Checkpointing Web3/Solidity REST APIs
Featured Work

Projects that push limits

Benchmark systems, AI agents, and 3D visualization—each project solving a real engineering challenge.

AI Research

TextCraft Agent Benchmark Harness

Reproduced the AgentGym-RL TextCraft environment to rigorously evaluate advanced LLMs including GPT-5.2. Built a production-grade local harness with automatic partial checkpointing, model-aware API request mapping, and action-level server log diagnostic tools for deep failure analysis.

Python GPT-5 API AgentGym-RL JSON Checkpointing LLM Evaluation
// Automated resume system handles long-running eval jobs without data loss
01
AI Systems

SWE-SEARCH — MCTS Software Agent

Implemented a software patch-generation agent powered by Monte Carlo Tree Search. Resolved complex environment bugs involving relative output paths and trajectory evaluations. Engineered a robust evaluation framework with absolute path resolution and full trajectory tracking.

Python MCTS SWE-bench Git Automation Testing
// Achieved a perfect 12/12 pass rate on the full test suite
02
Frontend Engineering

3D Remodeling Core UI @ VeraAI

Led the frontend migration from flat 2D interfaces to Three.js-powered 3D visualization at VeraAI Technologies. Transformed complex spatial AI data pipelines into fluid, real-time interactive environments for construction and remodeling workflows—bridging backend AI output directly to user-facing renderers.

Three.js React WebGL AI Pipelines 3D Spatial Data
// Pioneered 3D-first UI direction, promoted from intern to Frontend Lead
03
Career Timeline

Experience

A track record of rapid growth, from internships to leading frontier AI product development.

Frontend Lead 2024 – Present
VeraAI Technologies Inc.
  • Promoted from SDE Intern to Frontend Lead for demonstrating ownership and deep technical delivery
  • Led the full frontend migration to Three.js 3D visualization for construction and remodeling AI workflows
  • Bridged complex AI backend pipelines with intuitive, interactive user-facing interfaces
  • Drove cross-functional alignment between engineering and product teams on spatial data rendering
SDE Intern 2023 – 2024
VeraAI Technologies Inc.
  • Developed core frontend components integrating real-time AI pipeline data into interactive UIs
  • Rapidly self-taught Three.js to prototype the 3D visualization concept that became the product direction
  • Collaborated with AI researchers to translate model outputs into production UI patterns
MS — Information Systems 2022 – 2024
Northeastern University, Boston
  • Deep focus on AI systems, data engineering, and software architecture
  • Conducted AI benchmarking research across AgentGym-RL and MCTS-based agent frameworks
Hackathon Participant 2023
EasyA × Polkadot Hackathon — Harvard University
  • Built a Web3/smart contract project under the mentorship of Patrick Collins (Cyfrin)
  • Explored decentralized application architecture on the Polkadot ecosystem
✉️

Let's build something
remarkable

Open to roles in AI automation, agentic systems, and full-stack engineering. Let's connect.

Send a message
LinkedIn GitHub 🎓 Northeastern University