Hello, I'm Yashvardhan.
I build production AI systems, from multi agent pipelines to training, fine tuning and distilling small language models from scratch, turning research and complex workflows into reliable and automated software.
Small Language Model Specialist
I've distilled, fine-tuned, and trained SLMs from scratch — building compact, domain-specific models that rival much larger counterparts. From custom tokenizer design and pre-training on curated corpora to LoRA/QLoRA adaptation and knowledge distillation, I deliver lean models ready for production.
Featured Projects
Selected work from my AI engineering practice
Finetuning-Forge
End-to-end LLM fine-tuning pipeline with LoRA, QLoRA, and full fine-tuning. Experiment tracking, evaluation metrics, and production-ready model export.
Multi-Agent-Research-System
LangGraph-orchestrated framework routing queries to specialized AI agents that autonomously synthesize structured, evidence-backed research reports.
memory-infra
A standalone, budget-aware memory layer for LLM agents — extracts structured facts at write time and retrieves via hybrid vector, keyword, and entity-graph search. Up to 194× smaller prompts with bounded recall latency.
Work With Me
Available for freelance contracts and remote roles
Get in Touchyashvardhan49831@gmail.com