Anthony Adamo's Portfolio

My regular portfolio site.

recent posts

about

  • Stabilizing Context-Aware NPC Behavior Through Architectural Refinement

    This month’s goal was to stabilize and refine the Mineflayer NPC architecture after migration, with a focus on improving behavior reliability, dialogue consistency, and action recovery. While core features were already in place, the system needed stronger guarantees around repeatable behavior and clearer feedback loops during testing.

    To address this, I focused on consolidating the Behavior Manager, refining action sequencing, and improving how recent events and outcomes were summarized and passed to the backend language model. I also validated that NPC behavior and feedback persisted correctly across multiple test sessions, ensuring the system behaved consistently under repeated use.

    One challenge this month was identifying where decision failures originated when multiple actions were chained together. My strategy involved isolating each stage of execution to confirm that state was preserved correctly. This led to clearer separation of responsibilities and more predictable NPC responses during extended testing.

    Successes

    • Improved reliability of NPC behavior across repeated task executions.
    • Refined action sequencing and recovery logic.
    • Verified persistence of feedback and memory data between sessions.
    • Maintained consistent version control and documentation.

    Problems

    • Some recovery edge cases still appear under complex terrain conditions.
    • Debugging behaviors required more manual testing than expected.

    Going Forward

    • Expand automated behavior testing scenarios.
    • Improve fallback logic for navigation failures.
    • Begin collecting structured performance metrics earlier.

    Additional Reflection

    • Info Gained: I gained an understanding of behaviors, state persistence, and AI system reliability.
    • Course Application: Concepts from software architecture and AI systems informed the direct design decisions.
    • Preparation: With stability improved, I am ready to move into learning evaluation and performance analysis in the next phase.

  • Advancing the Dynamic Behavior and Feedback in My LLM Minecraft NPC

    This month’s goal was to advance the intelligence and responsiveness of my NPC agent by enhancing the action execution, integrating feedback, and stabilizing the system for milestone testing. I focused improving the communication between Minecraft and the backend language model, and validating the NPC behavior through targeted testing commands.

    To accomplish this, I worked within a Mineflayer architecture (originally starting with Fabric), implemented new command handlers, refined the recovery and action sequences, and profiled the performance using Clinic.js. This iteration strongly focused on accuracy, repeatability, and verifying the NPC’s ability to interpret tasks and adapt to player feedback.

    A core issue I faced early in the month was inconsistent decision responses from the NPC after executing several tasks. My strategy for solving this involved stepping through each component, such as commands, action handlers, dialogue routing, and queue logic, to isolate where state was dropping or being overwritten. This led to multiple refinements in the command structure and error handling, which resulted in more predictable behavior across sessions.

    Successes:

    • I successfully built and tested all the milestone commands for NPC evaluation.
    • The NPC is now far more stable during repeated test sessions.
    • Backend routing improvements increased the clarity and usefulness of LLM responses.
    • Version control and release management have been consistent throughout the month.

    Problems:

    • Several early errors in command handling caused ambiguous or repeated fallback responses.
    • Pathfinding tests initially produced inconsistent movement due to incomplete recovery logic.
    • The migration away from Fabric introduced work that delayed testing cycles.

    Going Forward:

    • Continue to expand the automated test scenarios to identify regressions early.
    • Maintain a higher update frequency with my advisor to validate feature direction.
    • Begin organizing performance metrics earlier in each milestone to reduce pressure.

    Additional Reflection

    • Time Utilization: I managed my tasks effectively, though some refactoring work took longer than expected due to the architectural changes.
    • Info Gained: This month improved my understanding of sequencing, backend routing to LLMs, and NPC state management.
    • Course Application: Concepts from AI programming, software architecture, and distributed systems were applicable during backend NPC communication.
    • Preparation: With the feedback and testing commands in place, I am positioned to begin the memory learning and evaluation cycles planned for the next milestone.

  • Early Progress on My Context-Aware NPC Capstone

    Hi, my name is Anthony Adamo, and I’m a Master’s student in Computer Science at Full Sail University. My capstone project is titled, “Context-Aware NPC Agent: Combining Retrieval-Augmented Generation with Environmental Learning in Minecraft.” I have a Bachelor’s Degree in Game Design, and I’ve always been interested in AI and interactive narratives. My long-term goal is to help advance AI systems that can comprehend morality and adapt in dynamic environments, for both for games and for diverse applications. This capstone represents a significant step in this journey.

    During the past month, my focus was on establishing the project’s foundation. I set up my development environment using IntelliJ IDEA for Java and integrated Fabric Loom to begin creating the Minecraft mod framework. I implemented a basic NPC entity that currently initializes within the game world to serve as the base for more complex behavior.

    Tools and Resources Used:

    • Java & Fabric Loom for Modding Support
    • Gradle for Build Automation
    • GitHub for Version Control
    • Reference Materials from the Minecraft Forge/Fabric Community Documentation

    Challenges: Getting the Gradle builds to compile correctly and resolving version mismatches between Fabric, Minecraft, and my IDE caused a few setbacks. After some troubleshooting, I resolved these by reverting to a stable project state and rebuilding my ideas from there.

    Successes:
    I successfully created a working project baseline and confirmed that my mod can run without breaking Minecraft. This was an important step because it gives me the assurance that the foundation is solid enough to build upon.

    Problems:
    The versioning issues and build failures cost me more time than expected. I realized I need to document each dependency and version number carefully so I don’t get swamped with debugging.

    Going Forward:
    I plan to keep a developer log that documents changes, dependency versions, and fixes for reference. This will make troubleshooting easier and also serve as part of my portfolio process documentation.

    Additional Reflection

    Time management was a bit off this month since I spent more hours troubleshooting than I planned. However, my previous coursework in Java programming, software design, and AI concepts supported my ability to navigate these challenges. My advisor interactions so far have helped me refine my project scope, and I’m preparing to move into the feature development phase where I’ll begin layering in environmental learning and retrieval-augmented generation for the NPC.

    This early foundation gives me confidence that I’m on the right course.