Shivam Choudhary
Verified Expert in Engineering
Artificial Intelligence (AI) Developer
Greater Noida, Uttar Pradesh, India
Toptal member since August 14, 2026
Shivam is a lead AI engineer and architect with 8+ years of experience building enterprise-grade AI platforms across renewable energy, IT Ops, procurement, and enterprise analytics. He specializes in agentic AI, RAG, conversational AI, LLM evaluation, and infrastructure. He combines strong foundations in software and data engineering with modern LLM technology to deliver production-ready AI solutions. Shivam helps clients modernize workflows, improve efficiency, and deliver measurable outcomes.
Portfolio
Experience
- Artificial Intelligence (AI) - 8 years
- Python - 7 years
- Data Engineering - 7 years
- Elasticsearch - 5 years
- Retrieval-augmented Generation (RAG) - 4 years
- Agentic AI - 3 years
- LLM Evaluation - 3 years
- Azure AI Foundry - 3 years
Preferred Environment
Visual Studio Code (VS Code), Claude, MacOS, GitHub, Docker
The most amazing...
...solution I've built for enterprise customers is an agentic AI platform that converts natural language into analytics and cuts insight time from days to minutes.
Work Experience
Lead AI Engineer | AI Solutions Architect
Global IT Services
- Designed and built a unified control plane architecture standardizing agent orchestration, tool routing, and AI governance across platforms, cutting new capability integration time by approximately 50% and enabling scalable, consistent rollout.
- Led the architecture and implementation of an enterprise-scale agentic AI analytics platform, converting natural language questions into structured analytics and database queries using RAG and semantic routing, achieving over 90% routing accuracy.
- Enabled self-service analytics for enterprise customers, reducing manual analyst effort by approximately 70% and cutting time to insight from days to seconds.
- Architected an LLM-powered insight generation and recommendation platform, reducing insight generation cycle time from multiple weeks to under one hour, delivering a +95% improvement.
- Built a conversational AI investigation assistant for multi-turn, context-aware queries, cutting time to resolution for operational investigations by roughly 60%.
- Developed a comprehensive LLM evaluation framework, integrating golden-dataset benchmarking, LLM-as-a-judge evaluation, and production monitoring to ensure consistent quality, safety, and performance across deployed AI applications.
Lead Software Development Engineer | Data Engineering
Global IT Services
- Led scalable data engineering solutions for enterprise analytics and AI platforms, covering high-volume data ingestion, transformation, enrichment, validation, and batch/near-real-time processing.
- Architected distributed ETL pipelines using Python, Elasticsearch, PostgreSQL, and Trino, implementing schema mapping, normalization, deduplication, validation, and query optimization.
- Built automated data quality frameworks for completeness, consistency, freshness, schema validation, and anomaly detection, improving the reliability of downstream analytics and AI workflows.
- Developed Python/FastAPI services and reusable data components, with CI/CD, Docker, Kubernetes, monitoring, logging, retry, and observability for production-scale data platforms.
Senior Software Development Engineer
IBM
- Architected and optimized enterprise software solutions using Python, improving application performance and reducing processing time by up to 30% across critical workflows.
- Designed and implemented scalable back-end services and REST APIs, integrating enterprise systems and external services while improving reliability, maintainability, and overall system performance.
- Collaborated with cross-functional engineering teams to troubleshoot production issues, automate operational processes, and deliver high-quality software solutions, reducing recurring manual effort by approximately 25%.
Site Reliability Engineer | Software Development Engineer
Vestas
- Developed and maintained real-time data processing pipelines using Apache Kafka, Apache Spark Streaming (PySpark), Amazon Kinesis, and PostgreSQL for large-scale data ingestion and processing.
- Implemented data extraction, transformation, enrichment, and schema mapping workflows to integrate streaming data with downstream applications and analytics platforms.
- Designed and built data quality monitoring solutions for both real-time and batch pipelines, capturing latency, completeness, volume, and aggregation metrics.
- Established end-to-end observability and operational monitoring using Amazon CloudWatch, Splunk, S3, Lambda, Redis, and Kinesis, enabling proactive issue detection and system reliability.
- Automated customer onboarding workflows by developing Python-based configuration and validation frameworks for assets, signals, and connectivity setup, reducing manual effort.
- Collaborated with cross-functional teams to improve platform scalability, streamline onboarding processes, and reduce customer onboarding time from four days to under two hours.
Experience
Natural Language Analytics and Agentic AI Platform
The platform integrates Elasticsearch/vector search, Azure OpenAI, and enterprise APIs to retrieve and synthesize relevant information. I also contributed to workflow state management, short-term conversational memory, evaluation frameworks, observability, and production-grade FastAPI services. The platform supports multiple enterprise customers and focuses on reliable, scalable, and explainable AI-driven analytics.
Enterprise Insight Generation and Recommendation Engine
I contributed to prompt engineering, LangGraph workflow design, evaluation strategies, observability, and integration with enterprise data services. Particular emphasis was placed on improving response accuracy, reducing irrelevant context, and building reliable AI workflows suitable for production enterprise environments.
Conversational AI Investigation Assistant
The system analyzes user questions, determines the appropriate investigation strategy, retrieves relevant information from enterprise knowledge and data sources, and produces contextual responses. I worked on workflow orchestration, short-term memory, query refinement, tool integration, prompt engineering, and evaluation of multi-turn conversations.
The solution was designed to maintain context across related questions while minimizing unnecessary context passed to downstream LLM calls, improving both response quality and efficiency.
Enterprise LLM Evaluation and LLM-as-a-judge Framework
AI Procurement Copilot – Multi-agent Vendor Proposal Analysis
https://bitwisethoughts.substack.com/p/multi-agent-ai-system-that-readsEducation
Master's Degree in Computer Applications
Vellore Institute of Technology (VIT) - Vellore, India
Bachelor's Degree in Statistics and Mathematics
Chaudhary Charan Singh University - India
Certifications
Microsoft Certified: Azure AI Engineer Associate
Microsoft
Google Cloud Professional Data Engineer
Google Cloud
Google Cloud Professional Machine Learning Engineer
Google Cloud
Skills
Libraries/APIs
Pydantic, Pandas, REST APIs, PyTorch, NumPy, Spark Streaming, Hugging Face Transformers, Asyncio, PySpark, Scikit-learn, Llama API, Claude API
Tools
Claude, GitHub, GitHub Copilot, Claude Code, Pytest, Splunk, Grafana, Kibana, Amazon CloudWatch, Logstash, Kafka Connect, Helm, Azure OpenAI Service, Claude Agent SDK, n8n
Languages
Python, SQL
Frameworks
LangGraph, Agentic Frameworks, Trino, Apache Spark
Paradigms
Model Context Protocol (MCP), Microservices
Platforms
Azure AI Search, Langfuse, CrewAI, Harness, Kubernetes, LangSmith, Azure, MacOS, Amazon Web Services (AWS), Apache Kafka, AWS Lambda, AWS IoT, Docker, Ollama, Visual Studio Code (VS Code)
Storage
Elasticsearch, Data Validation, PostgreSQL, Amazon S3 (AWS S3), Redis, MongoDB, Datadog, Data Lakes, Data Pipelines, Redis Cache
Other
FastAPI, LangChain, Azure Language Service, Sentence Transformers, OpenAI APIs, Google DLP, Agentic AI, Retrieval-augmented Generation (RAG), Context Engineering, LLM Fine-tuning, LLM Evaluation, AI Guardrails, Artificial Intelligence (AI), Data Engineering, DeepEval, ETL Pipelines, Persistence, Software Engineering, AI Chatbots, AI Agents, API Design, Error Handling, AI Harness Engineering, Observability, API Integration, RAG Systems, Anthropic, OpenAI, PDF, PDF Scraping, Evaluation, AI Agent Orchestration, Workflow Automation, LLM Integration, RAG Architecture, Model Evaluation, Back-end, Agentic AI Systems, Distributed Systems, Data Analytics, Communication, Attention to Detail, Code Review, Written Communication, Team Leadership, GitHub Actions, Recommendation Systems, LoRa, Azure AI Foundry, CI/CD Pipelines, Statistics, Mathematics, Open-source LLMs, Large Language Models (LLMs), Microsoft Azure, Screeners, Gemini, OpenAI SDK, ChromaDB, RAGAS, Data Science, Natural Language Processing (NLP), Machine Learning, Predictive Modeling, Data Processing Systems (DPS), Data Security, Data Transformation, Cloud Storage, Deep Learning, Amazon Kinesis, Cloud Computing, Data Quality, Real-time Data Pipelines, Multi-agent Orchestration, AI Governance, Prompt Engineering, ONNX Runtime, Conversational AI, Machine Learning Operations (MLOps), Vector Databases, Large Language Model Operations (LLMOps), Generative Artificial Intelligence (GenAI), A2A, AI Copilots, Optical Character Recognition (OCR), Directed Acrylic Graphs (DAG), Azure AI Document Intelligence
How to Work with Toptal
Toptal matches you directly with global industry experts from our network in hours—not weeks or months.
Share your needs
Choose your talent
Start your risk-free talent trial
Top talent is in high demand.
Start hiring