SS
Back to Projects

AI Research Orchestrator / Multi-Agent Research Scientist

Developed an autonomous multi-agent research system that plans literature reviews, retrieves academic papers, synthesizes findings, and generates structured research reports.

Node.jsTypeScriptSemantic ScholarOpenAlexLLMsAgent Orchestration

SYSTEM ARCHITECTURE

SYSTEM ARCHITECTUREDistributed AI research orchestrationResearch QuestionProblem intakeResearch PlannerQuestion-aware planLiterature RetrievalOpenAlex / Semantic ScholarResearch AgentsAnalyze evidenceExperiment PlannerMetrics + training planDistributed WorkersDMLF / PyTorch DDPResearch BriefEvidence + resultsOffline-demo mode uses deterministic agents and simulated workers where external services are unavailable.

Offline-demo mode uses deterministic agents and simulated workers where external services are unavailable.

Problem

Research workflows are fragmented across search, ranking, reading, and synthesis. The goal was to coordinate those steps into a repeatable agent-driven system.

Architecture

Planner, analyst, architect, critic, and consensus stages coordinate paper retrieval, citation analysis, literature summaries, and structured report generation.

Challenges

  • API rate limiting
  • Paper ranking
  • Agent orchestration
  • Tool abstraction
  • Offline execution mode

Benchmarks

Built research planning, paper retrieval, citation analysis, structured reports, and literature summary flows using external academic data sources and LLM-driven orchestration.

Lessons Learned

  • Specialized agents make research workflows easier to reason about
  • Consensus steps are useful when synthesis needs to be grounded in sources
  • Tool boundaries matter when mixing retrieval, ranking, and writing