Software engineer in Hyderabad building AI and multi-agent applications and the backend systems that keep them reliable at scale.
At Apxor I work with enterprise client teams like GCash, Porter and ICICI Bank, whose apps together serve over 100 million users.
- AI and agents: LLM agents that call tools to finish multi-step tasks, RAG over company documents, LLM-as-judge pipelines with evidence checked against the source
- Trustworthy AI: grounding checks, labelled eval sets that block regressions in CI, one model gateway with token caps, cost logging and rate limits
- Backend: rebuilt legacy services as Go microservices, 20% faster with 30% lower CPU and memory cost; event pipelines on Kafka and ClickHouse
- Platform: Kubernetes rollouts through GitOps (ArgoCD, Helm, Tanka), on-call and root-cause fixes
Latte · AI sales-call compliance and coaching · live Transcribes each speaker, scores every call against a compliance rubric with an LLM judge, and sends breaches to QA with timestamped evidence. Quotes are checked against the transcript, and CI blocks any change that misses a planted breach. REST API, signed webhooks, SDKs in TypeScript, Python, Go and Java.
Nemo · Multi-agent assistant over your own documents Switch between OpenAI, Anthropic and Gemini at runtime, with tool-calling agents behind a Go gateway. Retrieval pipeline, config-driven model factory, Kafka events, Temporal workflows, Kubernetes and ArgoCD delivery.
AI: LLM APIs (OpenAI, Anthropic, Gemini, Groq), LangChain, LangGraph, RAG, embeddings, Chroma, Whisper, LLM evals Backend: Go, TypeScript, Python, Java, FastAPI, Node.js, gRPC, Kafka, Temporal Data: PostgreSQL, ClickHouse, Redis, MongoDB Platform: Kubernetes, Helm, ArgoCD, Tanka, Terraform, AWS, GitHub Actions, Prometheus, Grafana
AWS Cloud Practitioner · Red Hat Certified System Administrator (RHCSA) · CodeChef 3 star (1637)


