Samyak | Devpost

Inspiration

Remote teams struggle with task visibility, time accountability, and async coordination, especially when tools are fragmented across task managers, chat apps, and documentation platforms. Studies on distributed work show that lack of real-time context and poor task clarity are major contributors to productivity loss in remote teams.

Samyak was inspired by:

The rise of agentic AI systems that can reason, act, and assist proactively
The need for a single system that combines task tracking, productivity analytics, and conversational assistance
Research showing that voice and conversational interfaces reduce cognitive load compared to manual dashboards

The goal was to build a system where teams manage work naturally — by assigning tasks, tracking time, and talking to an AI agent that understands project context.

What I Learned

Building Samyak provided hands-on experience with:

Agentic system design — designing AI agents that assist rather than just respond
Task lifecycle modeling (creation → assignment → execution → completion)
Time-based productivity metrics, such as completion rates and effort distribution
Webhook-based integrations for real-time notifications
LLM orchestration using Gemini for document generation, reasoning, and strategy

I also learned how to balance automation vs human control, ensuring AI suggestions remain explainable and actionable.

How I Built the Project

Samyak is designed as a modular agent-driven system:

Task & Time Management

Users can create, assign, and prioritize tasks
Each task tracks:
- Status (pending, in-progress, completed)
- Priority
- Logged working hours per user
Time logs are aggregated to generate productivity insights

Productivity Analytics

Completion rate calculated as:

$$ \text{Completion Rate} = \frac{\text{Completed Tasks}}{\text{Total Tasks}} \times 100 $$

Time utilization reports show:
- Hours spent per task
- Hours spent per team member
- Over/under-allocation trends

AI Voice Agent (Gemini)

Integrated Gemini LLM to act as a voice-enabled agent
Capabilities include:
- Conversational task queries
- Auto-generating documents and summaries
- Strategy suggestions based on task data
The agent operates with project context, not isolated prompts