Taofeeq Ismail Otu
I’m a Senior Backend Engineer and AI/Agentic Systems builder with 5+ years of experience shipping production-grade services for financial and enterprise products serving 150,000+ users.
About
I’m a Senior Backend Engineer and AI/Agentic Systems builder with 5+ years of experience shipping production-grade services for financial and enterprise products serving 150,000+ users. I focus on scalable backend architecture in Python (Django, FastAPI) and secure multi-tenant APIs, with deep experience integrating PSP/payment systems and KYC/identity verification flows.
I also lead end-to-end AI implementations, including RAG pipelines, LangGraph-orchestrated agentic workflows, and MCP tool-calling layers—always with cost controls, observability, and PII-safe processing. Currently, I’m architecting Converge AI at Heir Holdings Technologies, a multi-tenant RAG + MCP tool-calling service, while also leading backend delivery for a UBA savings and investment platform.
- 7 years' experience
- Joined September 2026
Industry experience
- Financial Services
- Software & Internet
- Education
- Government
- Professional Services
- Computers & Electronics
Work experience· 9 roles
Proposed and led the end-to-end architecture of Converge AI, a multi-tenant RAG and agentic MCP tool-calling service for Converge ERP. Designed a multi-tenant ingestion pipeline (PDF/Word/Excel/HTML/public URLs) with pgvector-namespaced embeddings and hybrid dense + BM25 retrieval with reranking and grounded, cited responses. Architected an MCP tool-calling layer for LangGraph workflows with slot-filling conversations, explicit confirmation gates, and RBAC-scoped Converge API execution (LLM does not access the database directly). Built cost optimization and observability (tiktoken budgeting, per-tenant/user throttling, caching, LangSmith/OpenTelemetry tracing). Implemented PII detection and redaction (Presidio + Nigeria BVN/NIN recognizers), with chat state split across Postgres (durable history/tool audit) and Redis (LangGraph checkpointing).
Designed and built an LLM-integrated AI interview system using an OpenAI SDK behind a model-agnostic FastAPI service layer for pluggable providers. Engineered an interview context pipeline where candidate resume content drives tailored interview questions and response handling. Implemented PII redaction ahead of every LLM call to ensure sensitive candidate data is removed before model usage. Delivered LLM-call orchestration (prompt design, context assembly, response handling) with structured logging and robust error handling.
Led backend architecture and delivery of FInTree (Monitree), a savings and investment app for UBA, acting as a primary owner for system decisions and shipping. Integrated SAML 2.0 and Okta for enterprise SSO. Built a production FastAPI multi-tenant bill-payments platform across UBA VAS markets with tenant isolation via API keys/JWT middleware. Implemented Flutterwave v4 payment lifecycle (cards, MoMo, virtual accounts, webhooks, refunds) and UBA fulfillment via Celery; also delivered secure onboarding flows (OTP, OAuth, idle timeout, token blacklist, step-up auth, device tracking, idempotency, digital signatures). Integrated UBA VAS SOAP (biller catalog, initialize/validate, fees, static/dynamic biller types). Added rate limiting, replay protection, request-integrity checksums, security headers, audit logging, SQLAdmin ops UI, tenant admin APIs, biller sync (scheduled/manual), in-app notifications, and SendGrid transactional email. Built multi-tenant KYC onboarding (FastAPI/PostgreSQL/
Education
Certifications and awards
Hire Taofeeq Ismail Otu today
Post a job and invite Taofeeq, or contact them directly. No hidden commission.