Performance · Engineered
Software Engineer · AI Agent Builder · Electronic Music Lover · Athlete
Engineer by discipline. Builder by obsession. I work on AI-powered systems—agents that reason, adapt, and scale. Hybrid athlete across tennis, golf, soccer, lifting, and running. Music lover. Firm believer that the best products—like the best sets—are engineered to make you feel something.
I'm a software engineer obsessed with building things that feel effortless to use but are surgically precise underneath. My work spans full-stack development—React, Next.js, TypeScript— and an increasingly deep focus on AI agents and agentic workflows. I think about AI not as a feature to bolt on but as a new substrate for building software altogether.
My engineering philosophy borrows from systems thinking: every component has a purpose, every interface a contract. I built HybridFit as a full-stack training tracker with NextAuth, Zod validation, and Vercel-optimised performance, and redesigned ESMON with a cohesive JavaFX design system that replaced visual chaos with hierarchy and intent. Both projects share a conviction: great software is inevitable once the mental model is right.
Outside the terminal I'm a hybrid athlete—half-marathon training, tennis, golf, soccer, lifting. I treat runs the same way I treat hard engineering problems: show up, stay systematic, trust the process. I also spend a worrying amount of time on logic puzzles and electronic music. These aren't hobbies; they're how I stay sharp.
Lifting weights. Squats, deadlifts, bench press, overhead press, etc.
Pattern recognition, constraint solving, systematic play.
Agents that reason, plan, and act—not just generate.
Clarity as a proxy for competence. Signal over noise.
Every system is a network of components.
Dec 2025 — Present
Software Engineer II
Java, Spring Boot, JUnit, Jenkins, GitHub Copilot
Built and shipped core platform features for an operations management system used across North America and Europe.
Led European operations rollout by executing production data loads, coordinating across 3 teams to ensure zero data gaps at launch.
Led cross-team release of major features, aligning dependencies across 3 teams to deliver on schedule with zero regressions.
Drove adoption of AI-assisted development (GitHub Copilot), standardizing development workflows across the team.
Jun 2024 — Dec 2025
Software Engineer
Redis, Java, Spring Boot, Angular, TypeScript, Jenkins
Introduced a Redis caching layer to surface real-time UI states, eliminating data inconsistencies across operations.
Built scalable Java/Spring Boot APIs for bulk CSV ingestion, enabling operations teams across 25+ European countries to onboard facility data.
Developed an Angular-based scheduling calendar UI used by facility managers for daily planning and operational monitoring.
Notes on building AI agents, engineering decisions, and lessons from shipping software.
Each project is a system—designed from first principles, built with precision, and shipped with care about every detail that the user will never notice.
AI-Powered Twitter Discovery & Engagement Agent
Personal productivity agent that searches Twitter for conversations worth joining, drafts replies that match your voice, and routes everything through Slack for approval before posting. The hard parts: orchestrating a multi-stage pipeline that pauses mid-execution for human decisions, enforcing cost controls before every external call, and teaching the model to write like you without claiming expertise you don't have.
Indian Railway Analytics Desktop App
Analytics desktop app for Indian Railways that turns raw binary journey data into filterable reports, time-series charts, and multi-section PDF exports. The hard parts: parsing undocumented binary formats, keeping filter state consistent across tabs without coupling the controllers, and rendering charts offscreen to prevent JavaFX deadlocks during PDF generation.
Full-Stack Fitness Platform for Hybrid Athletes
Training platform for hybrid athletes who don't fit neatly into one sport. The design challenge was making a complex system—multi-plan enrollment, flexible scheduling, granular workout logging—feel simple to use daily. Performance was non-negotiable: response caching and DB projections cut query load by over 99%. AI coaching recommendations are next.
© 2026 Tanish Nahata