CogitoNTNU/jarvis
Oct 2024 – Mar 2025
4 Months active
Languages Used
PythonJavaScriptYAML
Technical Skills
API IntegrationAgent DevelopmentBackend DevelopmentBug FixingCode OptimizationEvent Handling
Jon Bergland
PROFILE
Jon Bergland
Worked on the CogitoNTNU/jarvis repository, delivering a series of upgrades to its AI-driven agent framework and tool integration. Focused on modularizing the graph agent architecture, refining prompt handling, and improving event routing to support scalable tool usage and more reliable LLM responses. Enhanced both backend and frontend components using Python and JavaScript, implementing features such as perplexity-based agent orchestration, expanded toolsets, and frontend debugging utilities. Addressed deployment reliability by resolving Docker Compose conflicts and persistent storage issues. Prioritized code cleanup, maintainability, and standardization, enabling faster feature delivery and more predictable behavior across graph-based and state-managed components.
Overall Statistics
Feature vs Bugs
78%Features
Repository Contributions
41Total
Bugs
2
Commits
41
Features
7
Lines of code
2,018
Activity Months4
Your Network
39 peopleShared Repositories
10
Work History
March 2025
2 Commits
Mar 1, 2025March 2025: Stabilized deployment and storage for CogitoNTNU/jarvis by addressing Docker Compose issues. Removed duplicate phoenix service to eliminate startup conflicts and corrected persistent storage mappings after a rebase to ensure ChromaDB and related services retain data reliably. These changes improved startup reliability, data integrity, and developer onboarding for jarvis deployments.
2 Commits
Mar 1, 2025March 2025: Stabilized deployment and storage for CogitoNTNU/jarvis by addressing Docker Compose issues. Removed duplicate phoenix service to eliminate startup conflicts and corrected persistent storage mappings after a rebase to ensure ChromaDB and related services retain data reliably. These changes improved startup reliability, data integrity, and developer onboarding for jarvis deployments.
March 2025
February 2025
8 Commits • 2 Features
Feb 1, 2025February 2025 performance summary for CogitoNTNU/jarvis. Focused on modernization, reliability, and code hygiene to enable faster, safer delivery of features and more predictable AI behavior in graph-based components.
February 2025
8 Commits • 2 Features
Feb 1, 2025February 2025 performance summary for CogitoNTNU/jarvis. Focused on modernization, reliability, and code hygiene to enable faster, safer delivery of features and more predictable AI behavior in graph-based components.
November 2024
13 Commits • 4 Features
Nov 1, 2024November 2024, CogitoNTNU/jarvis: Delivered key product improvements and tooling enhancements focused on reliability, UX, and automation. Highlights include LLM Output Formatting improvements to ensure tool usage prompts return only the selected option name; expansion of perplexity-based tools and agent orchestration (perplexity_agent, agent_decision state, routing, and tool deciders including calendar and other tools), with corresponding prompt updates; frontend debugging utility to retrieve all displayed chat messages; and toolset expansion exposing core utilities (add_tool, find_files, read_file, read_pdf). These changes improve user experience, make tool invocation more reliable, and broaden automated capabilities.
13 Commits • 4 Features
Nov 1, 2024November 2024, CogitoNTNU/jarvis: Delivered key product improvements and tooling enhancements focused on reliability, UX, and automation. Highlights include LLM Output Formatting improvements to ensure tool usage prompts return only the selected option name; expansion of perplexity-based tools and agent orchestration (perplexity_agent, agent_decision state, routing, and tool deciders including calendar and other tools), with corresponding prompt updates; frontend debugging utility to retrieve all displayed chat messages; and toolset expansion exposing core utilities (add_tool, find_files, read_file, read_pdf). These changes improve user experience, make tool invocation more reliable, and broaden automated capabilities.
November 2024
October 2024
18 Commits • 1 Features
Oct 1, 2024Concise monthly summary for CogitoNTNU/jarvis (Oct 2024): Delivered a major overhaul of the Jarvis Graph Agent and Tools integration, improving the tool decision flow, prompt handling, event routing, and overall response quality. Reinstated a stable graph agent after refactors and added modular nodes and a router to support diverse tool usage, enabling more scalable tool integration. Implemented several stability and correctness fixes, including ensuring LLM calls are properly invoked, frontend streaming delivery, and cleanup of initialization logic. These changes reduced tool-usage latency, enhanced reliability, and established a foundation for future tool expansions and performance improvements. Technologies/skills demonstrated include Graph Agent architecture, modular node design, prompt engineering, tool routing, LLM invocation management, and streaming frontend delivery for faster user feedback.
October 2024
18 Commits • 1 Features
Oct 1, 2024Concise monthly summary for CogitoNTNU/jarvis (Oct 2024): Delivered a major overhaul of the Jarvis Graph Agent and Tools integration, improving the tool decision flow, prompt handling, event routing, and overall response quality. Reinstated a stable graph agent after refactors and added modular nodes and a router to support diverse tool usage, enabling more scalable tool integration. Implemented several stability and correctness fixes, including ensuring LLM calls are properly invoked, frontend streaming delivery, and cleanup of initialization logic. These changes reduced tool-usage latency, enhanced reliability, and established a foundation for future tool expansions and performance improvements. Technologies/skills demonstrated include Graph Agent architecture, modular node design, prompt engineering, tool routing, LLM invocation management, and streaming frontend delivery for faster user feedback.
Activity
Loading activity data...
Quality Metrics
Correctness84.4%
Maintainability86.8%
Architecture83.0%
Performance76.0%
AI Usage39.0%
Skills & Technologies
Programming Languages
JavaScriptPythonYAML
Technical Skills
AI/MLAPI IntegrationAgent DevelopmentAgent FrameworksBackend DevelopmentBug FixingCode CleanupCode OptimizationCode OrganizationDevOpsDockerEvent HandlingFront End DevelopmentGraph AgentGraph Databases
Repositories Contributed To
Overview of all repositories you've contributed to across your timeline