Hasta Bahadur Chhetri (Bimal Chhetry)

• Install-Centric User Acquisition: Managed performance-driven campaigns with a strong focus on user acquisition, installs, CPI, conversion rates, retention, and downstream monetization, optimizing campaigns based on the quality and value of acquired users—not just volume.
• ROAS, ROI & Revenue Optimization: Analyzed campaign performance through ROAS, ROI, LTV, eCPM, CPI, and revenue metrics, connecting acquisition spend directly to monetization and business outcomes to maximize return on investment.
• IAP & Ad Monetization: Worked across In-App Purchases (IAP), AdMob, and advertising-driven monetization, understanding how acquisition strategy impacts revenue, LTV, and overall profitability.
• Playable Ad Production: Built interactive playable ads using PixiJS and Playable Factory, reducing CPI and increasing interactive engagement by giving users a hands-on preview of the product before installation.
• End-to-End Attribution & Analytics: Managed attribution and performance tracking across AppsFlyer, Google Analytics, and Firebase, monitoring the complete user lifecycle—from impression → install → activation → retention → monetization → ROAS.
• AI-Powered Creative Production: Scaled creative production using Higgsfield, Creatify, and OpenArt, contributing to a 40% increase in creative visibility while enabling faster creative testing and iteration.
• Multi-Platform Performance Marketing: Managed and optimized campaigns across Google Ads and Meta Business Suite, using deep-funnel reporting to evaluate CPI, ROAS, retention, revenue, and overall campaign profitability.

Tools: Unity Ads, AppLovin, Google Ads, Meta Business Suite, AppsFlyer, Singular, Defold JS, Playable Factory, Creatify, OpenArt, Figma, DaVinci Resolve

Collaborated with the UNDP Nepal team to support the “My Vote Matters” national campaign during the 2026 General Elections. Focused on leveraging digital platforms to increase voter awareness and combat electoral misinformation.

• Content Production: Produced high-impact digital content, including a 60–90 second video focused on the first-time voter experience to drive youth engagement.
• Information Integrity: Actively monitored and reported digital misinformation, ensuring the distribution of factual, non-partisan electoral information.
• Strategic Outreach: Supported the public information campaign by distributing multimedia assets across various digital channels to reach diverse demographics in Nepal.
• Stakeholder Collaboration: Coordinated with international project managers and local media professionals to align campaign messaging with UNDP transparency standards.

Skills: Digital Outreach · Content Strategy · Video Production · Misinformation Monitoring · Campaign Management

Designed web and mobile interfaces, brand identities, and product packaging. Created multimedia content and managed social media campaigns, boosting engagement and visibility by 30%.

Tools: Figma, Photoshop, Illustrator, DaVinci Resolve, Meta Business Suite

Portfolio Highlights: Nizi iOS · Nizi Android · Nizistore

Learned and contributed to app interface design and coding, creating wireframes, high-fidelity visuals, branding, and graphics. Produced marketing visuals, edited promotional videos, and assisted with social media management.

View Medicos App

15K+ downloads · Solo-developed · AdMob monetization

React Native · Expo · TypeScript · Redux · Firebase

Interactive math quiz game featuring multiple difficulty levels, timed challenges and progress tracking.

React Native · Expo · TypeScript · Redux · Firebase

6+ apps published across iOS & Android, including Loksewa Tayari, Engineering Mock Test, Mehndi Design: AR Try-On 2026 and Sudoku Master.

Contributed to mobile/web redesign, UI/UX improvements, performance testing and feature integration.

Tech: React Native · Expo · TypeScript · Redux · Firebase · MMKV · AdMob

Sole author · Preprint · Published as Hasta Bahadur Chhetri

Empirical comparison of five machine-learning models (Random Forest, Decision Tree, Logistic Regression, Autoencoder, VAE) on CICIDS2017 (Random Forest F1=0.9965) with cross-dataset revalidation on Farm-flow, plus an autonomous Q-learning recovery layer (68.3% MTTR reduction) evaluated under compute and connectivity constraints. Tests four literature assumptions as falsifiable claims; two fail.

Topics: Machine Learning · Intrusion Detection · IoT Security · Reinforcement Learning · Python, PyTorch, scikit-learn

Sole author · Preprint · Published as Hasta Bahadur Chhetri

PRISMA-informed structured review of 26 selected studies (23 peer-reviewed + 3 preprints) on agentic AI for autonomous cyber-defense across ICS, IoT, and smart agriculture, classifying techniques against the five functions of the NIST Cybersecurity Framework.

Topics: Agentic AI · LLM Multi-Agent Systems · Critical Infrastructure Security · NIST CSF

Verified by Elsevier on ORCID (0009-0009-8909-2389)

Reviewed a full-length research article on Agentic AI, Decentralized Physical Infrastructure Networks (DePIN), edge intelligence, and blockchain governance for IoT hardware deployment.