// DATA ENGINEER & AI/ML ENGINEER

Yasaswin
Palukuri

Building autonomous AI that works while I sleep. Six agents, seven models, zero cloud bills — self-hosted on bare metal in Toronto.

Python LangGraph Ollama AWS Docker
yasaswin@m720q: ~/groot
$ groot status
6 agents running
7 models loaded (27GB)
6 ChromaDB collections
Ragas faithfulness: 1.000
Uptime: 4805 hours
$

The Project That Never Sleeps

A self-hosted multi-agent AI system running 24/7 on a Lenovo M720Q in Toronto.

Groot — Supervisor
qwen2.5:7b · routes every request
▼ ▼ ▼
Einstein
research
Tony
coding
Siva
tutor
Hybrid RAG · ChromaDB
BM25 (40%) + Dense (60%) retrieval
guardrails-ai · Ragas evals
Slack · Telegram · n8n workflows
7
Ollama Models
6
ChromaDB Collections
1.000
Faithfulness Score
24/7
Uptime
6
Active Agents
4805hrs
Running

RAG Pipeline Evaluation Results

Every retrieval pipeline in Groot ships with eval scores. Here are the real numbers.

ragas.evaluate(dataset)
# hybrid retrieval · production dataset · no cherry-picking
Faithfulness
0.000
Answer Relevancy
0.000
Overall Score
0.000
Evaluated with qwen2.5:7b as judge · nomic-embed-text embeddings · Ragas 0.1.21

How Groot Was Built

From a blank Ubuntu install to a fully autonomous 6-agent AI system. Every milestone is real.

FEB 2026

Conception and Planning

Selected Lenovo M720Q (i5-8500T, 16GB). Researched LangGraph, Ollama, ChromaDB. Decided on local-first, zero cloud AI dependency.

FEB 2026

Bare Metal Foundation

Ubuntu 22.04.5 LTS bare metal. LUKS full-disk encryption. WiFi driver (RTL8852BU). Tailscale VPN. First Ollama model running.

MAR 2026

First Agent Loop

FastAPI agent service. Groot supervisor routing to qwen2.5:7b. ChromaDB setup with 6 collections. Basic RAG pipeline working end-to-end.

APR 2026

Hybrid RAG and Load Scheduler

Custom asyncio priority queue with RAM headroom validation. Hybrid RAG: BM25(40%) plus Dense(60%). Real retrieval quality improvement measured.

MAY 2026

Job Search Pipeline

Adzuna Canada API. ATS-optimized resume tailoring. PDF generation via reportlab. Telegram delivery in under 60 seconds.

JUN 2026

Security and Evaluation

guardrails-ai: PII detection, prompt injection blocking, toxic content filter. Ragas evaluation: Faithfulness 1.000. 20-test red-teaming framework.

JUL 2026 · NOW

Full Autonomous Operation

6 agents live. 7 models, 27GB. n8n workflows. Notion nightly diary. React dashboard with 12 pages. Supervisor loop runs hourly. 4,190+ hours and counting.

Things I've Shipped

Production-minded projects — tested, evaluated, and mostly still running.

Skills That Ship

AI/ML & LLMOps
LangGraph LangChain Ollama ChromaDB FastAPI MLflow SHAP Scikit-learn Ragas guardrails-ai Streamlit
Data Engineering
Python SQL PySpark Spark Kafka Hadoop S3 Glue Lambda Redshift ETL/ELT
DevOps & Visualization
Docker Git Pytest Ruff CI/CD Linux Power BI Tableau Tailscale

Experience

Data Engineer @ Verita Network (Volunteer)
May 2026 – Present · Toronto

Building AWS ETL pipelines across S3, Glue, Lambda, and Redshift. Optimizing SQL workloads, monitoring pipeline health, and shipping in Agile sprints.

S3 Glue Lambda Redshift SQL
LLMOps Engineer @ Groot Project
Jan 2026 – Present · Toronto

Designed and operate a 6-agent AI system with hybrid RAG and thermal-aware load scheduling. Enforcing output quality with guardrails-ai, measuring it with Ragas, and automating everything else with n8n.

LangGraph Hybrid RAG guardrails-ai Ragas n8n
Credit Risk Scoring with XAI
May – Jun 2025

Built an explainable credit-risk model with SHAP and Scikit-learn, improving accuracy by 40% through feature engineering and WoE binning.

SHAP Scikit-learn WoE binning

Education

Lambton College
Graduate Certificate — Artificial Intelligence & Machine Learning
Toronto · 2025
KL University
B.Tech (Honors) — Computer Science
India · 2023

"If you wanna break it, break it till you build it!!"

Open to Data Engineer · ML Engineer · LLMOps Engineer roles