Principal Data Scientist | AI Engineer | GenAI Systems Architect

I design and build enterprise-grade AI systems — not experiments.
15+ years of experience delivering production-ready GenAI, ML, and optimization platforms across regulated, large-scale environments.

My work spans LLMOps, multi-agent systems, optimization engines, AI governance, and decision intelligence — turning cutting-edge research into real business impact.

• Architect LLM-powered enterprise platforms using GPT-4o, O1-mini, Llama-3
• Design multi-agent AI systems for knowledge synchronization, compliance, and decision support
• Build LLMOps & MLOps pipelines with governance, auditability, and monitoring
• Engineer hybrid optimization engines (OR-Tools + NSGA-II) for real-world constraints
• Translate ambiguous business problems into explainable, scalable AI solutions

Multi-agent GenAI system for regulatory document synchronization and compliance alignment.

Architecture

• Reader Agent → Comparator Agent → Derivative Agent
• Azure OpenAI + LangChain
• Semantic chunking, embeddings, traceable derivative generation

Impact

• Accelerated regulatory alignment
• Reduced manual compliance effort
• Full traceability & audit readiness

Enterprise solution for document analysis, comparison, and insight extraction.

Capabilities

• Context-aware summarization
• Entity extraction
• Intelligent document differencing
• Prompt pipelines + evaluation workflows

Tech GPT-4o, O1-mini, LangChain, Azure OpenAI

Conversational analytics platform enabling Text-to-SQL at enterprise scale.

Use Case

• Natural language querying across 125+ websites
• Sentiment analysis & insights via chat interface

Stack Azure OpenAI, Microsoft Fabric, SQL Server, Dashboards

Hybrid optimization + AI explanation system for asset utilization.

Core

• OR-Tools + custom NSGA-II (DEAP)
• Pareto analysis & constraint modeling
• XAI Decision Cockpit
• LLM-based AI Analyst for executive summaries

Outcome

• Improved asset utilization
• Cost reduction via optimized routing

GenAI & ML

• GPT-4o, O1-mini, Llama-3
• Prompt Engineering, Fine-tuning, Knowledge Distillation
• Multi-Agent Systems, Assistant Task Engineering

Architecture & Engineering

• LLMOps, MLOps, CI/CD (Azure DevOps)
• FastAPI Microservices
• API Design, System Orchestration

Optimization & Algorithms

• Multi-objective optimization
• OR-Tools, NSGA-II (DEAP)
• VRP, Pareto analysis, constraint modeling

Data & Infra

• Azure OpenAI, Microsoft Fabric
• Vector DBs, Embeddings
• SQL Server, Azure Blob

Tooling

• Python, FastAPI, Pandas
• LangChain, PyTorch
• React, Plotly, Folium

• Production > Proof of Concept
• Explainability is non-negotiable
• AI systems must be observable, governable, and resilient
• LLMs are components — not solutions by themselves
• Optimization + AI > AI alone

• Mentor engineers on AI/ML, GenAI systems, and career growth
• Write and speak about applied AI, optimization, and enterprise GenAI
• Focus on building thinking engineers, not tool operators

• LinkedIn: https://www.linkedin.com/in/tushitdave/
• Medium: https://medium.com/@tushitdavergtu
• Email: tushitdavergtu@gmail.com