Looking for a job

Rohit Yadav

@rohityadav15

Message

I build and train scalable LLMs, agentic AI systems, and ML infrastructure.

India

Message

What I'm looking for

I'm looking to build production-scale AI systems, especially LLM training, inference, agents, and ML infrastructure, while continuing hands-on research in distributed systems and open-source model development.

At AutonomousX, I’ve built and open-sourced 25+ TPU-trained LLMs from 120M to 1.4B parameters, training on 500B+ tokens and earning Google TPU Research Cloud support. I also develop LLM pipelines at Molsys, spanning pretraining, SFT, distillation, alignment, efficient inference, and Amazon Bedrock deployment.

I build reproducible AI systems beyond language models, including a clinical SNV ranker that reached 0.9186 ROC-AUC and a genomic pipeline that cut 2.5TB processing from 150+ hours to under 20. I also contribute to Meta LLaMA and Google DeepMind Gemma and build systems projects such as AI agents, compilers, and virtual machines.

Experience

Work history, roles, and key accomplishments

Current

AI Research Intern

Current

Molsys Pvt. Ltd.

Jun 2026 - Present (3 months)

Developing and optimizing LLM pipelines spanning pretraining, supervised fine-tuning, model distillation, RLHF-inspired alignment, and efficient inference. Building agentic AI systems and researching distributed training, inference optimization, quantization, and production-scale LLM deployment using Amazon Bedrock.

Large Language Models (LLM) Pretraining Supervised FineTuning Model Distillation RLHF Inference Optimization Quantization Amazon Bedrock Agentic AI Distributed Training

Current

Research Lead

Current

AutonomousX

Apr 2026 - Present (5 months)

Received Google TPU Research Cloud support for large-scale LLM training and research. Trained 25+ LLMs (120M–1.4B parameters) on 500B+ tokens using TPU pods ranging from v4-8 to v4-32 and v6e-64.

LLM Training TPU Jax Distributed Training Scaling Laws

Current

Open Source Contributor

Current

Meta LLaMA

Mar 2026 - Present (6 months)

Contributed to production-scale LLM repositories by fixing critical bugs and improving usability. Resolved tokenizer normalization and jax typing issues in Google DeepMind Gemma, and fixed incorrect CLI import causing ModuleNotFoundError in Meta LLaMA download pipeline.

LLM Tokenizer Jax Python Bug fixing

Selected Cohort Participant

Amazon

Jun 2026 - Jul 2026 (1 month)

Selected among 3,000 participants from 134,000+ applicants for Amazon's flagship ML Summer School 2026. Completed advanced training in Machine Learning, Generative AI, LLMs, and ML Systems, building hands-on projects focused on model optimization and production-scale AI.

Machine Learning Generative AI LLM ML Systems Model Optimization

Full-Stack Web Developer Intern

Skylark Express Delhi Pvt Ltd

Jun 2025 - Jul 2025 (1 month)

Developed an internal logistics dashboard to streamline operations and improve workflow visibility. Enhanced UI responsiveness and optimized application performance for day-to-day business operations.

Fullstack Development Dashboard Development UI UX Performance Optimization