anirudh@systems:~$ cat spec.yml focus: - Custom ReAct agent loops with deterministic state machines & self-correction - Asynchronous distributed AI execution & cloud orchestration (EKS, Celery, Redis) - Multi-provider LLM orchestration (Groq LLaMA 3, NVIDIA NIM, Gemini) - Sub-200ms latency retrieval (pgvector) & async microservices (FastAPI) benchmarks: reliability: "41 self-corrections triggered; 0/10 unresolvable failures (5/10 → 0/10 errors)" latency: "sub-200ms resume scoring API; <2s voice interview feedback" scale: "1,230+ records ingested across 3 sources with 518 automated merges" throughput: "447 live tech jobs ingested & scored (avg match 19.1, peak 77.4)"
| System | Core Architecture | Verified Metric | Artifacts |
|---|---|---|---|
| AgentKube | Kubernetes multi-agent execution platform · FastAPI Celery Redis AWS EKS Terraform Argo CD |
Distributed Async Jobs | Code |
| Hirenix | AI career intelligence platform · FastAPI pgvector Next.js NVIDIA NIM Groq |
sub-200ms API |
Live App • Code |
| Self-Correcting Agent | Deterministic ReAct loop & runtime self-correction engine | 41 self-corrections |
Code |
| AI Career CoPilot | Agentic job discovery, semantic matching & outreach pipeline · Gemini Celery Redis |
447 jobs scored |
Code |
| MessyData | Entity resolution, deduplication & golden record data pipeline · RapidFuzz SQLAlchemy |
518 auto-merges |
Code |
| PashuSwasthya | Offline edge ML livestock disease & breed classification · TFLite MobileNetV3 Flutter |
Edge Inference | Code |
Python • FastAPI • PostgreSQL / pgvector • Supabase • Docker • Kubernetes (EKS) • AWS • Redis / Celery • Terraform • Next.js / TS • Groq • NVIDIA NIM • TensorFlow Lite
Open to AI/ML & Full-Stack AI engineering roles • [email protected]