Dr. Shad Ahammed

• Led end-to-end AI product development from concept to deployment, integrating hardware and software for edge AI applications • Deployed and optimized TinyML speech recognition models on ESP32 microcontrollers, achieving real-time inference under strict memory constraints. • Designed and implemented Hardware-in-the-Loop test framework for automated AI model validation • Architected GNN-based sentiment scoring pipeline for quantitative trading signals with walk-forward retraining, and RAG-LLM inference

• Currently building anomaly detection models for industrial maintenance systems using sensor fusion and time-series analysis • Setting up RAG-LLM pipeline for maintenance recommendations - integrating domain-specific technical documentation with RAG • Designing Explainable AI framework to make anomaly detection outputs interpretable • Establishing project infrastructure and research roadmap for multi-sensor data integration, model explainability benchmarks, and LLM-based knowledge retrieval workflows

• Developed CNN and YOLO-based object detection models on CARLA synthetic datasets, achieving reliable obstacle and road-sign detection under simulated sensor drift conditions. • Designed and implemented image augmentation pipeline emulating real-world drift to stress-test model robustness before field deployment • Performed STPA safety assessments for safety-critical traffic scenarios, identifying control structure vulnerabilities in perception-to-decision pipelines

• Applied machine learning to emergency patient data for predictive health analysis, integrating FPGA and Arduino-based embedded systems to run deep learning inference at the edge in resource-constrained rescue environments. • Built LLM-based clinical decision support using Llama and BERT models, delivering context-aware recommendations from unstructured medical data to assist triage and treatment planning. • Authored and secured funding for BMBF scientific research proposals

• Acceptance Test Specifications for state-of-the-art integrated modular avionics equipments following /DO-254/, /DO-297/, AIRBUS standard • Developed general purpose automatic test equipment to perform Production acceptance tests for aerospace equipments • Define system requirements to test wired (Discrete/Analog/PWM) and BUS communication (CAN/A429/A664/A717) interfaces • Requirement engineering using DOORS document management

• Authored Acceptance Test Specifications for Integrated Modular Avionics systems in compliance with DO-254, DO-297, and AIRBUS standards • Developed general-purpose automated test equipment for aerospace production acceptance testing • Defined technical test requirements for wired and wireless avionics interfaces including CAN, ARINC 429, ARINC 664, and ARINC 717 protocols • Managed requirements engineering workflows using IBM DOORS to maintain traceability across specification, design, and test phases

• Reduced 75% of test run time by developing unique MATLAB based automated system for processing scattered protein droplet image • Successfully implemented image processing technique to deduce complex refractive index measurement

• Designed Python based machine learning model for detecting epileptic seizure with accuracy over 98% • Developed hardware prototype with FPGA(Altera) using VHDL without time slack • Designed test bench for verification of the software and debug with ModelSim

• Worked with IMU(MPU9150) sensor to obtain data from wireless netwrok • Designed sensor based electronic system with Raspberry Pi using OCTAVE/MATLAB

 Led several projects with efficiency-based planning from production till delivery  3 times customer credibility award in 1.5 years for maintaining project timeline  Increased production efficiency by 25-30% in 1 year by using smart planning tools e.g. SAP, ETS, FAST React

Artificial Intelligence in Healthcare and Automotive

1. Quantitative Business Analysis 2. Supply Chain

1. Electrical Circuit 1 and 2 2. Electronic 3. Electronic circuit simulation 3. Engineering Electromagnetics 4. Electrical Machine