AI Engineer | Backend Engineer | Systems Programmer

I build applied AI products, backend systems, and low-level Linux software. Most of my work sits around retrieval and agent workflows, distributed services, and systems close to the kernel and networking stack.

Portfolio • LinkedIn • Email

• Applied AI: LangGraph FAISS pgvector RAG Pipelines Agents RAGAS Langfuse MCP Google Gemini
• Backend: Go Python TypeScript FastAPI PostgreSQL Redis MinIO
• Infra & Ops: Docker Kubernetes Prometheus Grafana Alertmanager Ansible Caddy
• Systems & Runtime: C++20 C Linux Internals eBPF/XDP libbpf Namespaces cgroups

• Anchor • Python, FastAPI, LangGraph, PostgreSQL, pgvector, Langfuse
  Citation-grounded RAG system over Indian financial regulations using hybrid retrieval, reranking, and continuous evaluation.

• Steward • Python, LangGraph, MCP, E2B, PostgreSQL, pgvector
  Autonomous GitHub issue triage and resolution agent with explicit state-machine control, sandboxed tool execution, and memory across runs.

• Photon • Go, Postgres, Redis, Kubernetes
  Async image-processing backend with worker queues, object storage, and observability.

• Guardian • Python, Prometheus, Ansible
  Monitoring and auto-remediation pipeline that bridges Prometheus alerts to automated Ansible playbooks.

• K-Watch • C++, eBPF/XDP
  High-performance network observability engine with SYN-flood detection and auto-mitigation built on kernel-level instrumentation with sub-5% CPU overhead.

• Vortex • C++, Linux
  Zero-dependency container runtime built from scratch using Linux namespaces, cgroups, and filesystem isolation.

• PacketForge • C++20
  Userspace L2/L3 network stack that hand-crafts ARP, ICMP, and raw Ethernet frames over AF_PACKET.

ECE student at NIT Silchar, currently seeking internships and software engineering roles.