01
2026 / Python
Documentation Assistant
RAG application that indexes technical documentation URLs and answers questions with retrieved context.
YASH KHAIRNAR / PERSONAL OPERATING MANUALVER. 2026.10AVAILABLE FOR NEW GRAD 2027
01INTRODUCTIONAI / ML ENGINEER
SAN JOSE, CA
SAN JOSE, CA
AI / ML
ENGINEER.
Learning, reasoning, and perception.
I'm an MSCS student at San José State University, graduating in May 2027. My current research explores action-conditioned world models, and my industry experience includes deploying computer vision systems and optimizing inference.
I'm interested in how agents learn unfamiliar environments, how language models are trained and evaluated, and how visual representations support intelligent behavior.
YK
CS / AI
01—26
CS / AI
01—26
02STATEMENT
I build and study systems that learn from data, experience, and perception.
My work sits between research and engineering: make the idea legible, make the system useful, then learn from what breaks.
DOWNLOAD RÉSUMÉFOCUS AREASWHAT I EXPLORE
01World Models & Reinforcement Learning
Learning environment dynamics, exploration, and planning.
02Language Model Training & Evaluation
Understanding the pipeline from data and training to adaptation, evaluation, and inference.
03Computer Vision & Inference
Building and deploying visual systems efficiently.
03WORK INDEXSELECTED BUILDS / 2025—26
Selected
systems.
Four ways of making intelligence useful: retrieve it, translate it, coordinate it, deploy it.
02
2026 / Next.js
Nook
Cafe discovery platform with personalized recommendations, location-aware search, and a mobile/web product surface.
03
2025 / PyTorch
ThinkFlow
EEG-to-text translation system using VQ-VAE, BART Transformer, and LSTM Seq2Seq with attention.
04
2025 / AWS Lambda
Sentinel
Serverless email marketing platform with analytics, AI-assisted content generation, and multi-region AWS deployment.
04FIELD LOG
A timeline of
useful mistakes.
Graduate Student Researcher
San Jose State University
Working on a JEPA-inspired, action-conditioned learner world model that predicts student knowledge-state transitions from programming trajectories and supports adaptive instructional planning.
Graduate Mentor | AI4ALL
Stanford University
Mentored 30+ students in AI/ML concepts, Python programming, and project development through hands-on workshops and technical mentorship.
AI/ML Engineer
Accurate Industrial Controls Pvt. Ltd.
Engineered an end-to-end predictive maintenance system for industrial generators using anomaly detection and RUL prediction models, achieving 93% accuracy across 500+ hours of telemetry data.
AI Intern
Accurate Industrial Controls Pvt. Ltd.
Built an anomaly detection system for copper coil inspection using PatchCore and YOLO, integrating object tracking and image compression to improve inference efficiency.
Deep Learning Intern
ResoluteAI Software
Developed a facial recognition attendance system using MTCNN for detection and a custom ANN classifier for embeddings, achieving 90% accuracy with less than 2% FPR.
Notes from
the bench.
Explanations, comparisons, and first-principles thinking about the systems I build.
01 / ESSAY
Transformers: The Idea That Changed AI Forever
Deep dive into the transformer architecture that revolutionized machine learning, from attention mechanisms to practical implementations.
READ 02 / ESSAYNeural Machine Translation: The Mathematics of Attention
Exploring how the RNNSearch mechanism solved the information bottleneck in sequence-to-sequence models for translation tasks.
READ 03 / ESSAYPerformance Analysis: LSTM vs RNN
Comparative study using Google Speech Commands dataset. Understanding the trade-offs between architectural choices in recurrent networks.
READ06FINAL PAGE