Shivam Yadav
Building intelligent, scalable, and reliable AI systems for real-world applications.
AI/ML Engineer + GenAI, RAG, AI Agents, and MCP
About Me
I'm an AI/ML Engineer with experience in building scalable, reliable, and production-ready software systems for real-world applications. My work spans AI/ML, backend development, intelligent automation, and modern data-driven systems, with a focus on creating practical solutions that are efficient, impactful, and built for deployment.
I enjoy solving problems across AI products, backend services, information retrieval, workflows, and system design using modern cloud, data, and application technologies. Outside of work, I also enjoy playing football, which reflects my interest in teamwork, discipline, and consistency.
Skills
Tools I reach for every day
Projects
Systems I've designed & shipped
Experience
Building production AI systems at scale
Associate Software Engineer (AI/ML)
Jul 2024 — Present- Built AI-powered solutions using Semantic RAG, Azure Cognitive Search, NLP, and Multilingual AI for legal and education-focused use cases, improving retrieval efficiency by ~60% and learning outcomes by ~40%.
- Developed production-grade AI/ML and ETL pipelines with Python, FastAPI, PostgreSQL, ADF, Synapse, Databricks, and Purview for scalable and governed deployment.
- Designed monitoring dashboards using Elasticsearch and OpenSearch to track performance and operational metrics for 460K+ users.
- Built a multimodal RAG system for law enforcement applications, supporting audio, video, image, and document intelligence with real-time identification across 100K+ records.
PythonFastAPIGenerative AIAI AgentsRAGMultimodal AIVector SearchOCRAzureAWSDatabricksElasticsearchOpenSearchKibanaStudent Technical Intern (AI/ML)
Jan 2024 — Jul 2024- Enhanced ingestion and retrieval workflows using Azure Form Recognizer, OCR, semantic chunking, and open-source LLMs.
- Optimized inference pipelines with LLaMA, Mistral, and Ollama, reducing model inference cost by ~40%.
Azure Form RecognizerOCRSemantic ChunkingLLaMAMistralOllamaPython
Verified credentials in AI & engineering
Certificate of completion: Claude 101
Anthropic
Claude Certified Architect — Foundations
Anthropic
Microsoft Certified: Azure AI Engineer Associate
Microsoft
Aisera AI Workflows (Intermediate Level)
Aisera Academy
Aisera GPT (Basic Level)
Aisera Academy
Entrepreneurial Management
Great Learning
Articles
Long-form pieces on AI systems, agents & infrastructure
Your AI Has Amnesia. Redis Iris Just Fixed the Most Annoying Problem in AI
LLMs have no built-in memory — every message looks like a first interaction. This piece breaks down how Redis Iris's unified context layer (Agent Memory, LangCache, Context Retriever) plus semantic caching gives AI agents real memory while cutting redundant token costs.
Claude Code Is Powerful. Without These 7 Skills, It's Also Unreliable
Seven practical skills — Skill Creator, Superpowers, GSD, review commands, Context Mode, Claude Mem, and Frontend Design — that turn Claude Code from a capable-but-inconsistent assistant into a reliable, production-grade coding partner.
Get in touch
Email, social, or the contact form