Ajinkya Mawal

Work

OnFinance AI Peak XV Surge · $4.2M

Forward Deployed Engineer & AI Engineer · Mumbai · May 2026 – present

OnFinance builds agentic AI for compliance, risk and audit in banking and financial services, deployed on-prem or in the client's private cloud. I'm on the side of the company that goes into the client's environment and makes it work there.

  • Own enterprise deployments end to end: scoping with client stakeholders, building, rolling out into production inside their environment, and iterating after launch.
  • Build and maintain the agent harness behind those deployments: orchestration loops, tool interfaces, context assembly, and deterministic fallback paths that keep outputs auditable.
  • Design multi-agent workflows that automate regulatory compliance audits, financial risk modelling, and accounting operations.
  • Integrate on-prem and hybrid AI pipelines into legacy core banking systems, working alongside the client's engineering, security, and risk teams.
  • Run continuous evaluation and tracing against live deployments, gate releases on regression checks, and hold latency and reliability budgets in production.

Multi-agent orchestration · tool calling · RAG · evaluation harnesses · tracing · on-prem and bare-metal · AWS · FastAPI · PostgreSQL · Redis · Docker · Kubernetes

Karta $3M seed · Oct 2026

AI Product Intern · Bengaluru · Jun 2025 – Aug 2025

Karta builds voice AI agents that handle inbound customer conversations for banks, lenders and insurers. It raised a $3M seed round led by PeerCapital and Cedar Hill Capital in October 2026.

  • Built an AI observability platform on LangSmith traces and agent run evaluations to monitor token latency, flag regressions, and cut customer escalations.
  • Built a prompt-based automation engine and standardised SOPs, reducing customer onboarding time by 60% and improving CX accuracy by 10%.
  • Optimised multi-step agent workflows and fallback routing to scale support resolution throughput.

LangSmith · LangChain · evals · prompt automation

Swiggy

Data Science Intern · Bengaluru · Apr 2025 – Jun 2025
  • Built an enterprise RAG chatbot over internal operational knowledge bases for self-serve retrieval across teams.
  • Architected an automated candidate-screening pipeline for HR, modernising resume ingestion and recruitment.

RAG · vector databases · Python · SQL


Full details in the resume.