Open to opportunities

Vaibhavi Madhav Deshpande

@vaibhavimadhavdeshpa

Message

AI Engineer building production multi-agent systems for real-time decisioning.

Zimbabwe

Message

What I'm looking for

I’m looking to build reliable, low-latency AI agent systems that ship to production. I want a team that values safety, observability, and measurable improvements in cost, accuracy, and real-time decision quality.

I’m an AI Engineer focused on turning LLM and agent ideas into production systems that are fast, reliable, and measurable. Most recently, I built and deployed a production-grade multi-agent AI platform for real-time crypto recommendation, portfolio monitoring, and autonomous risk analysis.

I design resilient agent workflows with provider failover, structured validation (Pydantic), and event-driven orchestration using async Python, Redis pub/sub, and streaming. I also optimize inference cost and scalability through selective LLM invocation, staged retrieval, batching, and caching—while implementing safety and execution controls like confidence gating, human-in-the-loop safeguards, and real-time monitoring/alerting.

Experience

Work history, roles, and key accomplishments

Current

AI Engineer - Multi-Agent

Current

Allyzent

Jan 2026 - Present (5 months)

Built and deployed a production-grade multi-agent AI platform for real-time crypto recommendation, portfolio monitoring, and autonomous risk analysis using FastAPI, Python, PostgreSQL, and Redis. Engineered LLM-powered agents with provider failover, low-latency event-driven orchestration, and AI safety controls (confidence gating, rule overrides, human-in-the-loop) while optimizing inference cost

fastAPI Python PostgreSQL Redis Multi Agent Systems AI Agents Pydantic Prompt Engineering

Software Engineer (Backend)

Business & Engineering Solutions Team Inc. (BEST)

Jun 2025 - Dec 2025 (6 months)

Designed and implemented Kafka-based event-driven backend systems to process real-time ERP sensor and operational data streams, reducing reporting latency by 50%. Built and deployed streaming-data AI agent systems that consumed Kafka events to evaluate system state and trigger autonomous operational decisions, enabling continuous forecast recalculation and adaptive scheduling.

Kafka Event Driven Architecture Streaming Pipelines Distributed Systems Real time Data Processing AI Agents Backend Engineering Python System Design Performance Optimization

Research Assistant - ML

University of Central Florida

Aug 2024 - May 2025 (9 months)

Performed exploratory data analysis and statistical hypothesis testing on clinical datasets to identify key patterns and validate model assumptions, improving feature selection and model robustness. Developed and tuned supervised ML models with a 15% RMSE reduction and built a RAG-based conversational system using HuggingFace embeddings and Pinecone vector similarity search to improve top-k retrie

Hypothesis Testing Hyperparameter Tuning Model Validation RAG Pinecone Similarity Search Clinical Data Analysis