Asad Karim-Senior AI Developer
Experience
Senior AI Developer
Neuland.ai AG
- Architected and deployed a production-scale GraphRAG system using Neo4j, embeddings, and multi-hop reasoning over 120M+ nodes, improving answer precision by 32%, reducing hallucinations by 41%, and lowering retrieval latency by 38%.
- Designed and implemented an enterprise agent ecosystem using Model Context Protocol (MCP), exposing internal APIs, databases, and services as secure callable tools for autonomous workflows and system integration.
- Designed and deployed a production LLM-based email routing agent using Microsoft Graph API, MCP, and Azure OpenAI, achieving 96% routing accuracy, reducing manual triage workload by 65%, and decreasing response times from 18 hours to under 4 hours.
- Implemented autonomous agent self-correction pipelines using iterative feedback loops (Ralph Wiggum), enabling reliable error detection, automated remediation, and production-safe execution.
- Developed a multimodal semantic search platform using multimodal LLMs and vector embeddings, enabling semantic discovery across 250k+ image and video assets and improving search recall by 48%.
Senior Data Scientist
Sparrow Networks GmbH
- Benchmarked data with BERT, LLaMa 3, Mistral 7B for semantic search and anomaly detection, boosting classification accuracy by 18%.
- Applied knowledge distillation to compress large transformer models, reducing inference latency by 45% and cloud costs by 30% while retaining 92% accuracy.
- Fine-tuned LLMs via PEFT/QLoRA across 15+ client datasets, reducing memory usage by 40% and improving predictive maintenance forecasts by 17%.
- Deployed RLHF-optimized chatbot for spare-part classification, lowering misclassification from 12% to 4%.
- Operationalized agentic frameworks (LangGraph) for RAG-based recommendations, improving precision by 28%.
Machine Learning Engineer
Sparrow Networks GmbH
- Built an automated scraping pipeline for manufacturer and supplier inventories, reducing manual scraping time by 90%.
- Migrated production database from PostgreSQL to Google BigQuery for efficient model generation and reporting.
- Created automated Change Data Capture (CDC) pipeline using Kafka, Debezium and Spark.
- Automated client reporting via Google Cloud Functions, minimizing manual effort.
- Implemented data cleaning pipelines with Apache Airflow for better orchestration.
- Designed a text pattern recognition pipeline using OpenAI’s Retrieval-Augmented Generation (RAG) and Pinecone vector database to detect anomalies in manufacturer data.
- Developed an OCR pipeline leveraging a multi-modal LLM (Vertex AI) to extract text from scanned documents, improving the customer journey of Sparrow app and reducing human errors to nearly 0%.
Professional Data Scientist
CarOnSale (Castle Tech GmbH)
- Migrated to a recommendation system using latent matrix factorization, improving recommendation quality by nearly 15%.
- Oversaw A/B testing to refine recommendation algorithms, focusing on buyer clusters and timeframes.
- Developed an internal Alpaca-based LLM for efficient Snowflake query generation for stakeholders, reducing data retrieval time from nearly 5 hours to within 30 minutes.
- Contributed to the Market Price Prediction project, applying regularization techniques to reduce prediction errors from 18% to 11%.
- Assumed managerial responsibilities including roadmap planning, team workflow optimization, and Jira management following Agile (Scrumban) practices.
Junior Data Scientist
CarOnSale (Castle Tech GmbH)
- Improved a LightGBM-based recommendation engine using advanced feature engineering, boosting performance by 6%.
- Optimized model architectures, increasing mean average precision and coverage metrics by over 20%.
- Co-developed a used-vehicle market pricing model, increasing pricing accuracy and contributing to a 10% rise in customer retention.
- Developed explainability endpoints for pricing predictions using SHAP.
Junior Data Engineer
CarOnSale (Castle Tech GmbH)
- Transitioned into a data engineering role during capacity constraints, supporting both data science and BI teams.
- Managed data pipelines and ETL workflows ensuring high availability for analytics and reporting.
- Developed an automated Slack bot using AWS Lambda and SendGrid, reducing campaign execution time from 2 hours to 5 minutes.
- Created Scrapy spiders for competitive intelligence, reducing market analysis turnaround time by 80%.
- Led migration of legacy workloads to Apache Spark, achieving a 3x increase in processing speed and 2x overall system efficiency.
Werkstudent Data Science & Engineering
CarOnSale (Castle Tech GmbH)
- Developed end-to-end ML prototypes, ETL pipelines, and BI integrations supporting production data teams across recommendation systems and computer vision use cases.
- Implemented PII redaction for auction images using TensorFlow and Keras, reducing manual inspection time from 2 hours to 10 minutes.
Werkstudent System Administrator
Enway GmbH
- Supported enterprise Windows and Linux environments, resolving system issues and automating operational workflows using Ansible.
- Improved system reliability through automated backups and updates, reducing manual intervention by 50%.
Data Scientist
VentureDive
- Engineered an R-based analytics platform for oncology clinical trials using dplyr, ggplot2, and Shiny, including a custom oncology-focused R library.
- Improved scalability by containerizing analytics services with Docker and deploying to AWS S3.
- Integrated Python and R workflows using Rpy2, and developed REST-style analytics endpoints in JupyterLab.
Industry Experience
See where this freelancer has spent most of their professional time.
Experienced in Information Technology, Automotive, Healthcare, and Pharmaceutical.
Business Area Experience
See which departments and functions this freelancer has contributed to most.
Experienced in Information Technology, Business Intelligence, Product Development, Research and Development, and Project Management.
Skills
Languages
- Python, R, Sql, Java, C, Javascript, Ruby, Sap-abap, Rust
Llm & Nlp
- Bert, Llama, Mistral, Gpt (Openai), Vertex Ai, Hugging Face Transformers, Rag, Graphrag, Neo4j, Pinecone, Langchain, Langgraph, Crewai, Mcp, Lora/qlora, Rlhf
Data & Ml
- Tensorflow, Keras, Pytorch, Lightgbm, Polars, Scikit-learn, Knime
Mlops / Orchestration
- Apache Airflow, N8n, Docker, Jenkins, Travisci, Github Actions, Aws (Lambda, Sagemaker, Ec2, S3), Gcp (Bigquery, Cloud Functions), Azure, Snowflake
Databases
- Postgresql, Mongodb, Bigquery, Pinecone
Knowledge Graphs
- Neo4j, Entity Linking, Ontology Modeling, Graph Traversal, Golden Answer Systems
Developer Tools
- Pycharm, Vscode, Jira, Heroku, Ansible, Spark, Hadoop
Languages
Education
Otto-von-Guericke University Magdeburg
Master of Science in Data and Knowledge Engineering · Data and Knowledge Engineering · Magdeburg, Germany
Institute of Business Administration
Bachelor of Science in Computer Science · Computer Science · Karachi, Pakistan
Statistics
Experience
Global Experience
Expertise
Qualifications
Profile
Frequently asked questions
Do you have questions? Here you can find further information.
Where is Asad based?
What languages does Asad speak?
How many years of experience does Asad have?
What roles would Asad be best suited for?
What is Asad's latest experience?
What companies has Asad worked for in recent years?
Which industries is Asad most experienced in?
Which business areas is Asad most experienced in?
Which industries has Asad worked in recently?
Which business areas has Asad worked in recently?
What is Asad's education?
What is the availability of Asad?
What is the rate of Asad?
How to hire Asad?
Average rates for similar positions
Rates are based on recent contracts and do not include FRATCH margin.
Similar Freelancers
Discover other experts with similar qualifications and experience
Experts recently working on similar projects
Freelancers with hands-on experience in comparable project as a Senior AI Developer
Nearby freelancers
Professionals working in or nearby Magdeburg, Germany
Most recent projects
FRATCH works with many companies and recruitment agencies. Here you will find our recently posted projects and opportunities.