I build autonomous AI systems that replace manual work— not decks.
Nizaal Khot · AI Engineer specializing in Agentic AI, MCP, and Multi-Agent systems. Shipping production autonomy at enterprise scale.
A live multi-agent run, right here.
Pick a prompt, hit run. Three agents collaborate — Research scouts, Analyst synthesizes, Writer composes. Status chips show live state, output streams in.
Scan sources, collect citations
Synthesize and rank findings
Compose final output
Run the pipeline to see the streamed report appear here.
Senior AI engineer. Production agents, not demos.
3.5+ years of software engineering, now focused on autonomous multi-agent systems and generative AI at enterprise scale. I architect orchestration graphs in LangGraph, build MCP tool servers, fine-tune open-source models with LoRA/QLoRA, and ship RAG pipelines that clear evals — not just vibes.
Six surfaces where I operate.
Every card is a practice I ship in production — not a keyword.
Agentic AI
Designing multi-agent ecosystems where specialized agents collaborate through orchestration graphs.
Model Context Protocol
Building and integrating MCP tool-calling servers that let LLMs interact with real systems.
Vibe Coding
Agent-first, AI-augmented development where the human directs and agents execute.
Multi-Agent Systems
Hierarchical, supervisor, and swarm patterns in LangGraph and CrewAI.
RAG at Scale
Hybrid search, re-ranking, and evaluation pipelines for enterprise knowledge.
LLM Ops
Fine-tuning, evaluation (RAGAS / LangSmith), cost & latency optimization.
What I reach for.
A focused stack for agentic systems — optimized for evals, observability, and real latency in production.
Agentic Frameworks
06- LangChain
- LangGraph
- CrewAI
- Microsoft Agent Framework
- Google ADK
- AWS Strands
Generative AI
06- GPT-4 / 4.1
- Claude 4.5
- Gemini 2.5
- Hybrid RAG
- Re-ranking
- CoT · ReAct · DSPy
Protocols & Tooling
03- Model Context Protocol
- Tool-calling
- Function-calling
Machine Learning
04- PyTorch
- Hugging Face
- LoRA / QLoRA
- Scikit-learn
Data & Memory
05- Pinecone
- Weaviate
- Neo4j
- PostgreSQL
- Graphiti
LLMOps & Evals
05- RAGAS
- LangSmith
- AWS Bedrock
- SageMaker
- CI/CD for ML
Languages
06- Python
- Asyncio
- Pydantic
- FastAPI
- SQL
- JavaScript
Four years at Dentsu Global Services.
Each role compounded — from scraping pipelines to autonomous multi-agent ecosystems.
- Jul 2025 — Present
Senior Software Engineer (AI)
Dentsu Global Services
Mumbai, India
- Architected a hierarchical multi-agent ecosystem in LangGraph with 3 specialized agents — Code Generator, Reviewer, Feedback — reducing coding time by 86% via autonomous production-ready Python generation.
- Replaced brittle manual prompts with DSPy optimizers, improving enterprise compliance adherence by 28% and cutting token latency.
- Deployed a Multimodal Agent on AWS Bedrock for handwritten document ingestion using multi-model ensemble approaches — reducing manual data entry by 52%.
- Engineered an autonomous end-to-end QA agent pipeline (LangGraph) that fully replicated QA engineer duties for email campaign validation, accelerating launch cycle times significantly.
- Mar 2024 — Jul 2025
Software Engineer
Dentsu Global Services
Thane, India
- Designed Generative AI pipelines using GPT / Gemini models to automate customer workflows and cut manual processing overhead.
- Engineered large-scale ETL pipelines ensuring data quality, integrity, and reliability across enterprise datasets.
- Mentored junior engineers on prompt engineering, MLOps, and model optimization best practices.
- May 2023 — Mar 2024
Associate Software Engineer
Dentsu Global Services
Thane, India
- Conducted R&D on GitHub Copilot integration strategies, identifying workflow optimizations adopted across the engineering team.
- Built Python web-scraping pipelines (2.x / 3.x) for large-scale e-commerce data extraction.
- Delivered automated scraping solutions that cut human effort significantly across data operations.
- Collaborated in Agile sprints using JIRA for task tracking, sprint planning, and cross-team support.
- May 2022 — May 2023
Consultant
Dentsu Global Services
Thane, India
- Developed Python 2.x / 3.x web-crawling scripts for e-commerce platforms supporting data ops and cluster-based processing.
- Provided cross-team technical support and site analysis for data scraping and collection workflows.
Shipped systems, not prototypes.
Each project made it to production with measurable business impact.
Hierarchical Code Agent
Generator → Reviewer → Feedback
Production multi-agent system delivering autonomous Python generation with built-in review and compliance loops. 86% coding-time reduction in enterprise workflows.
- LangGraph
- Claude 4.5
- DSPy
Multimodal Ingestion Agent
Handwritten docs → structured data
Deployed on AWS Bedrock for handwritten document ingestion using a multi-model ensemble — reduced manual data entry by 52% across enterprise operations.
- AWS Bedrock
- Python
- Ensemble
QA Agent Pipeline
End-to-end email campaign QA
Autonomous agent pipeline that replicates QA engineer duties for email campaign validation — accelerating launch cycle times across marketing ops.
- LangGraph
- Python
- LangSmith
Certified. Schooled. Shipping.
AWS Certified: Machine Learning — Specialty
Amazon Web Services
credly.com
B.E. Information Technology
M.H. Saboo Siddik College of Engineering · Mumbai
2022 · CGPA 8.48 / 10
Let's build autonomy.
Open to senior AI engineer roles, agentic consulting, and research collaborations.
Based in Mumbai, India · Available worldwide (remote & hybrid) · +91 9594988780