CV
Data & BI Engineer background, AI/ML engineering work, and full career history.
Contact Information
| Name | Muhammad Awais |
| Professional Title | AI/ML Engineer |
| muhammadawais.de@gmail.com | |
| Location | Ilmenau, Thuringia |
Professional Summary
AI/ML Engineer building LLM and deep learning systems end to end — retrieval and agent architectures, neural audio models, and the evaluation infrastructure that measures whether either actually works. Three years of prior production data engineering (Razor Group Berlin, Teradata, NETSOL).
Experience
-
2023 - 2024 Berlin, Germany
Data Engineer
Razor Group GmbH
E-commerce operations and marketplace analytics for Amazon seller aggregation.
-
2023 - 2023 Islamabad, Pakistan
Data Engineer
Teradata
Data engineering in the telecom domain.
-
2020 - 2022 Islamabad, Pakistan
BI Engineer
NETSOL Technologies
Enterprise data warehousing in the leasing and finance domain.
Education
-
2024 - Present Ilmenau, Germany
M.Sc.
Technische Universität Ilmenau
Media Engineering
- Faculty of Electrical Engineering and Information Technology
- Thesis: neural audio codec research, supervised by Prof. Gerald Schuller (TU Ilmenau), co-supervised by Andreas Brendel (Fraunhofer IIS, Erlangen)
-
2016 - 2020 Faisalabad, Pakistan
B.Sc.
National Textile University
Software Engineering
- Grade: 1.9 (German grading scale)
Skills
Languages
Certificates
- Microsoft DP-900 (Azure Data Fundamentals) - Microsoft
- Teradata Vantage - Teradata
Projects
-
LogiMind
Multi-agent RAG over public logistics operational documentation. Fixed-orchestration agent pipeline (planner, retriever, responder) with hybrid BM25 and dense retrieval, Qdrant, FastAPI, Streamlit. 0.98 faithfulness, ~9.4s median latency, $0.028 per query. 108 tests, Dockerized, CI on every push, deployed live.
- AutoGen, LangChain, RAGAS, LangSmith, Qdrant, FastAPI, Streamlit, AWS
-
Neural Audio Codec (audio_cod)
Low-latency neural speech codec research for my M.Sc. thesis. Causal transformer encoder/decoder with residual vector quantisation, targeting 8-16 kbps for real-time speech. Eight-phase controlled curriculum; quality plateaued at 3 bits per codebook.
- Python, PyTorch, torchaudio
-
edc_pred
Predicting energy decay curves from room geometry. 103M-parameter model reaching R² = 0.9995 on T20 and C50 prediction, trained on measured spatial room impulse responses.
- Python, PyTorch
-
SRIR-Analysis
Comparative study of PCA, LDA, and EFA for room acoustics classification. Parameter sensitivity analysis across octave bands on the Stolz et al. (2024) SRIR dataset.
- MATLAB
-
Job Application Tracker
Self-built application tracker with AI-powered intake. React front end, Supabase backend, Claude API for parsing job postings.
- React, Supabase, Claude API