Hello, I'm Ankur. I'm an AI Engineer Associate at MULTIPLAI, building LLM, RAG and AI agent systems that make it to production.
About me
I'm an AI Engineer Associate at MULTIPLAI, where I design and build AI solutions for enterprise clients: from the first requirements workshop to the architecture to the service running in production.
Most of my work lives where LLMs meet real software: RAG pipelines, AI agents, workflow automation and the backend APIs that make them reliable. I hold a Master's in Engineering and Sustainable Technology Management from SRH Berlin University of Applied Sciences.
I like practical AI: systems people actually use, that are safe, measurable and maintainable.
My journey
AI Engineer Associate · MULTIPLAI GmbH
Germany · Full-time
Design and develop AI solutions for enterprise applications, from architecture to deployment. Turn client needs into AI architectures, and build backend services integrating LLMs, RAG pipelines, AI agents and workflow automation.
Apr 2026 – PresentAI Intern · MULTIPLAI GmbH
Germany
Started at MULTIPLAI as an intern on enterprise AI projects, which led into the full-time role.
Sep 2025 – Mar 2026Master's in Engineering and Sustainable Technology Management
SRH Berlin University of Applied Sciences · Berlin, Germany
Coursework: Machine Learning, Artificial Intelligence, Data Science, AI Development, Deep Learning, Natural Language Processing, Software Development, Statistics and Robotics.
Oct 2023 – Mar 2026Postgraduate Diploma in Advanced Computing (CDAC)
India
Specialised in software development and led a team project building a full-stack enterprise application.
2022 – 2023Bachelor's in Mechanical Engineering
India
Engineering fundamentals.
2017 – 2021My projects
Fruit Detection Robot
A simulated robot arm in Webots that spots ripe fruit with OpenCV colour segmentation and picks it.
- Computer vision
- Robot simulation
- Python
- OpenCV
- Webots
Predictive Maintenance Dashboard
Health monitoring for turbofan jet engines: predicts Remaining Useful Life from NASA CMAPSS sensor data and shows it in an interactive dashboard.
- NASA CMAPSS
- RUL prediction
- Python
- scikit-learn
- Random Forest
- Streamlit
Semantic Book Recommender
Describe a mood or a plot in plain words and get matching books, using LLM embeddings and semantic search over a book catalogue.
- Semantic search
- Emotion-aware
- LangChain
- OpenAI
- Hugging Face
- ChromaDB
Mental Health Copilot
A RAG assistant over real Reddit mental-health discussions: FAISS vector search, LLM answers and crisis detection with safe-response guardrails.
- RAG
- Crisis detection
- Responsible AI
- LLMs
- FAISS
- NLP
- MongoDB
- Reddit API
Tax-Law Reasoning Generator
Generates synthetic, multi-step tax-law cases with ground-truth reasoning chains to benchmark how well LLMs really reason. Built on the MuSR framework.
- LLM evaluation
- Chain-of-thought
- Python
- GenAI
- MuSR
Side quests
AI Job Application Tracker
A Gemini-powered agent that reads job emails from Gmail, extracts company, role and status, and keeps a Google Sheet up to date.
GitHub →AI-ML Mastery Handbook
A hands-on notebook tour: data visualisation, deep learning, transformer-based NLP, embedding clustering and more.
GitHub →
My skills
LLMs & GenAI
Proof: Mental Health Copilot, Book Recommender
- LLMs
- RAG
- AI agents
- LangChain
- OpenAI
- Gemini
- Hugging Face
- Prompt design
Vector search
Proof: Book Recommender, Mental Health Copilot
- ChromaDB
- FAISS
- Embeddings
Machine learning
Proof: Predictive Maintenance
- scikit-learn
- TensorFlow
- Keras
- Deep learning
- NLP
- Statistics
Data
Proof: Predictive Maintenance
- Python
- Pandas
- NumPy
- SQL
- PostgreSQL
- MongoDB
Backend & cloud
Proof: Work at MULTIPLAI
- REST APIs
- Flask
- Docker
- Azure
- Google Cloud
- Git
Frontend & apps
Proof: This town
- TypeScript
- React
- Next.js
- Tailwind
- Streamlit
- Three.js
Vision & robotics
Proof: Fruit Detection Robot
- OpenCV
- Webots
Contact me
Want to get in touch? Write here.