Hi, I'm Amna

I build AI systems that work in the real world and help teams ship them.

AI Engineer · Researcher · Fullstack

When I'm not training models, I'm raising cherry shrimp. 🦐

click to explore 🦐

// who I am

A bit about me

I'm an AI engineer and researcher who builds systems that work in the real world. My work spans machine learning, LLM-powered agents, full-stack AI products, and the research that underpins all of it.

I'm a partner at Rikolabs, an AI agency where I currently lead technical direction and delivery as acting tech lead and fractional CTO; working directly with clients, scoping projects, and making sure things actually ship.

I believe the best engineers bridge raw technical depth with human understanding. Da Vinci studied anatomy to paint better and physics to design bridges; I take the same approach. Published researcher in AI ethics, healthcare AI, and bias detection. Always learning, always building.

I write occasionally at m005h1.github.io — mostly thinking out loud.

When I'm not building

Chess

Shrimp keepingReadingSelf-growth

Amna Kausar

AI Engineer · Fullstack

10+Publications

3+Years in AI

Knowledge Base

Project Logs

AI Agency Partner // Rikolabs

[//] Rikolabs

Fractional CTO

Acting tech lead and fractional CTO at Rikolabs, an AI agency. Lead technical direction, scope and deliver AI projects, manage client relationships, and drive product from brief to deployment. Handling business development and delivery in parallel.

Tech LeadFractional CTOAI DeliveryClient Management

Live Feed

KayAI // Therapeutic AI System

[//] Kindpath

Lead AI Engineer

Contributed to the development of KayAI, a HIPAA-compliant therapeutic conversational AI enabling guided emotional de-escalation interventions at scale. Architected LLM-based intervention flows and REST API integration for mobile deployment.

AWSLLM AgentsREST APIHIPAA

Live Feed

Multimodal AI Content Detector

[//] WeCatchAI

content

Created multimodal AI agent that detects AI-generated content and verifies potential misinformation across text, image, and video formats.

OpenAIAzureHugging Face

Live Feed

GDM Tracking Platform

[//] Ajeenkya DY Patil University

Best Thesis Award

Led comprehensive women's health research on gestational diabetes mellitus (GDM), developing and implementing a full-stack platform with AI models to detect, monitor, and predict GDM progression. Built complete web application with backend API, database, and deployed machine learning models for clinical tracking and analysis.

React/Node.js (MERN)FastAPIPyTorchHugging FaceMongoDBMachine Learning Research

Read Log

Governance AI Agent

[//] MentalWealthAcademy

Onchain Summer Awards

Contributed to the core decision-making architecture of AzuraOS, a governance AI agent for structured community decision-making. Built on DeepSeek, n8n, and Supabase; deployed into a blockchain-based Farcaster mini-app. Reached 256 users within 48 hours of launch.

DeepSeekSupabasen8nBlockchain

Live Feed

Seagrass Mapping UNet

[//] Omdena - Italy Chapter

Volunteering

Mapped seagrass meadows using satellite imagery (Google Earth API, Sentinel-2). Developed UNet model of 81.2% accuracy and deployed via Streamlit.

TensorFlowGoogle ColabStreamlit

Source

Diabetes Meta-Classifier

[//] Samsung Innovation Campus

Prediction

Meta-classifier for diabetes prediction and risk factor analysis, leveraging Python and PyTorch for model development for the Samsung Innovation Campus capstone project.

PyTorchMatplotlib

Source

Knowledge Base

Research Logs

2026Conference paper

Random Forests and Ensemble Methods for Diabetes

Developed ensemble machine learning models (bagging, stacking, voting) using SVM, KNN, and Random Forest to predict diabetes, achieving higher accuracy than baseline classifiers through comprehensive data preprocessing and model integration.

Access Document

2025Conference paper