Reachy Mini Robotic Receptionist
A perception pipeline exploring how a robot can recognize visitor cues before deciding when to greet.
Tempe, Arizona · Mathematics (Statistics) @ ASU
Aarsh
Nandra.
Applied AI Engineer
& Researcher
I build AI systems, research infrastructure, and intelligent products—from multi-agent simulations and multimodal evaluation systems to full-stack AI applications.
Selected work
Things I've made
work in the real world.
Projects at the intersection of applied AI, research engineering, and product systems.
AI-Assisted Prosthetics
A photo-to-3D workflow that combines generative models and physical measurements to support prosthetic design.
InsuranceAI
AI-assisted life-insurance intake, preliminary estimates, life-policy document Q&A, and licensed-broker handoff.
EduCAST Evaluation
Evaluation infrastructure that turns instructional-video evidence into structured scoring and targeted refinement routes.
Dementia Assistant
An accessible mobile prototype combining conversational AI, caregiver profiles, and routines for dementia-support contexts.
VoxProof
A Chrome extension and AI backend that turn spoken responses into cryptographically signed, browser-verifiable text.
Child-Peer Simulation
A Concordia-based classroom simulation exploring student interactions and memory over time.
Research
Research is where I learn to ask better technical questions.
At ASU's DaRL Lab, I work on educational AI evaluation infrastructure and the emerging problem of multi-agent classroom simulation—taking seriously both what a system does and how we know it works.
EduCAST
Co-first author; AAAI manuscript in preparation. Multimodal evaluation pipelines for personalized educational-video generation.
Classroom simulation
Ongoing Concordia-based research into teacher-student interaction, interventions, and emergent patterns.
Engineering experience
Applied AI Engineer ASU Next Lab
Voice-first mobile AI prototype for dementia-caregiving research.
Undergraduate Researcher DaRL Lab · ASU
Multimodal evaluation and research infrastructure for educational AI.
AI Systems Engineer Ender-IT
Recruiter-facing matching and workflow infrastructure.
Project archive
More systems, experiments, and prototypes.
HiNC Local AI PDF Q&A
Fully local, multi-document semantic Q&A with page-level sources and no cloud AI pipeline.
AI / ML · Data · SystemsGitHub ↗
BlogPostAI
Modular autonomous pipeline that researches, synthesizes, styles, and renders long-form articles from one prompt.
AI / ML · SystemsGitHub ↗
HealthInsightEngine
Personal-health data analysis infrastructure using rule-based reasoning, statistics, and trend detection.
Data · AI / MLGitHub ↗
Ender-IT Recruiting Platform
Internally deployed recruiter workflow with resume parsing, candidate state, matching, retrieval, and job management.
Full Stack · Systems · AI / MLInternal deployment
CarbonIQ
Award-winning Solana prototype connecting wallet transactions, estimated carbon impact, and staking incentives.
Full Stack · Systems · DataGitHub ↗
Technical toolkit
Useful tools,
organized by the work.
AI & ML
LLM systems · multimodal AI · RAG · embeddings · vector search · Whisper · scikit-learn
Research & evaluation
experimental design · benchmarking · human evaluation · ablations · agent-based simulation · statistics
Product & backend
Python · FastAPI · REST APIs · SQLAlchemy · document processing · rate limiting
Frontend & mobile
React · React Native · Expo · TypeScript · Vite · Tailwind CSS
Data & infrastructure
SQL · PostgreSQL · Supabase · pandas · NumPy · synthetic datasets
Core languages
Python · JavaScript · TypeScript · Java · SQL
About
I like making complicated things concrete.
I'm a Mathematics (Statistics) student at Arizona State University's Barrett, The Honors College. I work across AI research and engineering because I enjoy understanding complex systems: how they behave, how they can be evaluated, and how to turn an ambiguous idea into software people can use.
I'm especially interested in intelligent systems, simulation, machine learning, and difficult technical problems that reward careful reasoning.