Resume_

cat ~/experience.md

Download PDF

# Profile

Lead AI & Data Engineer with 5+ years of experience designing, building, and operating production AI, Generative AI, data, and cloud systems. Combines AI architecture with hands-on implementation across agentic workflows, retrieval-augmented generation, model and tool routing, evaluation, observability, secure backend integration, and production improvement. Experienced in translating product and stakeholder needs into technical roadmaps, scalable architectures, measurable quality criteria, and maintainable releases.

Deepest current expertise spans LangChain, LangGraph, Langfuse, RAGAS, ReAct, multi-agent systems, hybrid retrieval, embeddings, vector databases, graph-database retrieval, semantic search, Python, FastAPI, MCP, Azure, Docker, and CI/CD. Leads a four-person team developing a RAG solution, mentors engineers, establishes architecture/testing/release standards, and remains hands-on with difficult design and delivery problems. Broader foundations include enterprise data platforms, machine learning, computer vision, deep learning, secure API platforms, and academic AI research.

# Impact

  • ▹ 94% retrieval accuracy — achieved on an internal benchmark by designing a hybrid RAG architecture that combined embeddings, keyword search, metadata filtering, and reranking.
  • ▹ 50% fewer agent errors — reduced production agent errors using RAGAS and Langfuse evaluations for relevance, groundedness, tool-call correctness, and response quality.
  • ▹ 60% faster integration — designed AI-MCP to standardize AI-to-tool and AI-to-data interactions across 15+ products, reducing retrieval and integration time.
  • ▹ 40% lower API costs — reduced query latency by 25% through model benchmarking, intent-based routing, and asynchronous FastAPI services.
  • ▹ €850,000 revenue — built governed Databricks pipelines that acquired more than 30 internal customers.
  • ▹ Accelerated component sourcing by 35% through LangGraph multi-agent workflows with ReAct-based tool selection for retrieval, validation, and ranking.
  • ▹ Delivered 10+ secure Azure services that saved 30+ hours weekly and supported 100+ customers.
  • ▹ Advised 15+ engineering teams on AI practices and supported framework adoption by 500+ developers.
  • ▹ Maintained 99.9% service uptime using Azure alerts and reduced critical interruptions and downtime by 35% year over year.

# Experience

Lead AI and Data Engineer

Carl Zeiss AG

July 2022 — Present

Technical Leadership, AI Architecture & Client Solutions

// AI Architecture, Agentic Systems, and RAG

  • ▹ Lead a four-person team developing a RAG solution; translate product priorities into technical roadmaps, mentor engineers, and define architecture, testing, and release standards.
  • ▹ Designed a hybrid RAG architecture using embeddings, keyword search, metadata filtering, and reranking; achieved 94% retrieval accuracy on an internal benchmark.
  • ▹ Built LangGraph multi-agent workflows with ReAct tool selection for retrieval, validation, and ranking; accelerated component sourcing by 35%.
  • ▹ Designed evaluation workflows with RAGAS and Langfuse for relevance, groundedness, tool-call correctness, and response quality; reduced agent errors by 50%.
  • ▹ Benchmarked models and implemented intent-based routing with asynchronous FastAPI services; reduced query latency by 25% and API costs by 40%.
  • ▹ Architected a production RAG implementation using Azure AI Search and OpenAI; increased response relevance by 50% and reduced hallucinations.
  • ▹ Tuned Azure AI Search indexes and vector data flows; reduced retrieval latency by 30%.

// AI-MCP Agentic RAG Platform

  • ▹ Designed AI-MCP to standardize AI-to-tool and AI-to-data interactions across 15+ products, reducing data retrieval and integration time by 60%.
  • ▹ Implemented the RAG layer as a staged component flow: agentic chunking prepares content for retrieval; hybrid search retrieves across vector and graph databases; reranking prioritizes retrieved context; and RAGAS evaluates the resulting RAG behavior.
  • ▹ Created the RAG layer to provide a reusable grounding and retrieval path for AI-MCP-connected products while keeping tool and data interactions consistent.

// Enterprise AI Platform, Security, and Delivery

  • ▹ Delivered 10+ secure Azure services using Python, FastAPI, MCP, Entra ID, Azure API Management, Docker, and CI/CD; saved 30+ hours weekly and supported 100+ customers.
  • ▹ Secured APIs through dual-layer Azure API Management and Entra ID authentication with least-privilege access controls.
  • ▹ Standardized authentication across 15+ products and reduced access-management overhead by 60%.
  • ▹ Maintained 99.9% uptime using Azure alerts and monitoring; reduced critical interruptions and downtime by 35% year over year.
  • ▹ Used Terraform across 20+ data products, reducing resource-deployment time by more than 40%.
  • ▹ Built a business dashboard for the Zeiss Global Customer Center with 30+ metrics and presented it to global stakeholders and the Zeiss CFO.

// Data and ML Platform Foundations

  • ▹ Led development of a master-data product spanning 7 business units and 15 countries.
  • ▹ Built governed Databricks ETL pipelines with Unity Catalog; generated €850,000 in revenue and acquired more than 30 internal customers.
  • ▹ Developed dbt transformations with reusable macros and schema-level tests.
  • ▹ Built CI/CD automation for Unity Catalog schema changes and orchestrated end-to-end Databricks ingestion and dbt transformation tasks.
  • ▹ Used MLflow on Databricks to support reliable, high-quality training data for enterprise ML models.
  • ▹ Led creation of the eVA 3.0 data platform using a data-mesh architecture with independently managed data products.

// Technical Leadership, Adoption, and Stakeholder Advising

  • ▹ Advised 15+ engineering teams on AI practices and supported framework adoption by 500+ developers.
  • ▹ Combined hands-on architecture and implementation with mentoring for a four-person RAG team; established shared standards for delivery and quality.
  • ▹ Translated product and stakeholder needs into roadmaps, architecture decisions, quality measures, and executable delivery work.
  • ▹ Communicated technical outcomes through customer-facing services, cross-business-unit delivery, and executive reporting.

Data Analyst

Regiocom Netzdienste

September 2021 — February 2022

Machine Learning, Computer Vision, and Energy Analytics

  • ▹ Built data-processing pipelines and machine-learning models for energy-demand prediction and optical character recognition.
  • ▹ Improved energy-demand prediction by 22% using a Python processing and modelling pipeline across 90+ dimensions with PCA.
  • ▹ Built real-time versus historical forecast visualizations using Seaborn and Matplotlib.
  • ▹ Achieved 96.73% handwriting-recognition accuracy for digitizing manual energy readings using OpenCV and Keras.
  • ▹ Implemented an RNN model with TensorFlow using handwriting samples with and without text separation.

Data Analyst

Deloitte Risk and Financial Advisory

January 2018 — September 2019

Data Analysis, Automation, Risk, and Client Reporting

  • ▹ Automated monitoring and reporting for distributed financial and operational data pipelines with Python; saved 20 engineering hours weekly.
  • ▹ Performed root-cause analysis on claims data; generated $10M in revenue and identified risks that saved 2% of organizational costs.
  • ▹ Reduced dynamic-loading latency by 1.2% by fragmenting data across distributed database nodes.
  • ▹ Reduced monthly reporting error rates by up to 80% through VBA and Excel Macro automation.
  • ▹ Received a client appreciation award after reducing anomaly-reporting time by 91.6% through automated retrieval, processing, and reporting.
  • ▹ Resolved data-quality and integration issues across structured and unstructured datasets.
  • ▹ Translated stakeholder requirements into validated data models, Python and SQL transformations, and operational workflows.

# Education

MSc. Data Engineering

Otto-Von-Guericke University

October 2019 — July 2022

Data engineering, machine learning, deep learning, computer vision, and medical-image analysis.

  • ▹ Focus areas included data engineering, machine learning, deep learning, computer vision, multimodal modelling, and medical-image analysis.
  • ▹ Completed the multimodal deepfake detection, VoxelMorph medical-image registration, and StyleGAN fingerprint-generation projects described above.

Bachelor of Engineering

Chandigarh University

August 2014 — May 2018

Graduated with a GPA of 8.52 in the top 1% of the cohort.

  • ▹ Graduated with a GPA of 8.52 in the top 1% of the cohort.
  • ▹ Built a Java application connected to a networked SQL database for automated exam-question generation, reducing preparation time by 40%.

# Skills

// Generative AI, LLMs, and Agentic Systems

LangChain LangGraph Langfuse RAGAS ReAct tool-using LLMs multi-agent systems model routing tool routing MCP prompt engineering prompt versioning guardrails groundedness testing tool-call evaluation response-quality evaluation AI observability

// RAG, Retrieval, and Search

Retrieval-augmented generation agentic chunking embeddings vector databases graph databases for retrieval vector search semantic search hybrid vector/graph-database retrieval keyword search metadata filtering reranking Azure AI Search RAGAS evaluation relevance measurement hallucination reduction

// Machine Learning, Deep Learning, and Computer Vision

Python PyTorch TensorFlow Keras OpenCV RNNs StyleGAN VoxelMorph multimodal fusion image registration PCA MLflow custom training loops in academic research benchmark evaluation model-quality measurement

// Backend, APIs, and Software Engineering

Python FastAPI asynchronous services REST APIs MCP integrations API design Git automated testing architecture standards release standards schema-level testing maintainable service integration

// Cloud, Security, Platform Engineering, and Delivery

Azure Azure AI Search Azure Functions Azure API Management Entra ID Azure DevOps Docker Terraform CI/CD least-privilege security dual-layer authentication alerts monitoring uptime management secure production delivery

// Data Engineering and Analytics

SQL Databricks Unity Catalog Delta Lake dbt MLflow ETL distributed data pipelines distributed databases master data data mesh schema tests structured and unstructured data data quality monitoring reporting MySQL Oracle DB2 MongoDB Seaborn Matplotlib executive dashboards

// Leadership, Product, and Communication

AI architecture technical roadmaps product-priority translation mentoring engineering standards framework adoption stakeholder advising customer-facing delivery cross-functional collaboration executive reporting technical communication end-to-end technical delivery

# Languages

  • ▹ English — Bilingual
  • ▹ German — Basic/A1
  • ▹ Hindi — Native