Akshay Sakariya
Open to work2 years experienceBackend Developer
Data Science Graduate | Applied Machine Learning
Standing Rank
Rank Not Available
Developer Badges
Skills & Technologies
39 skillsWork Experience
Data Engineer
Bosch Engineering GmbH
- Deployed daily Airflow ETL workflows on a Kubernetes-hosted infrastructure integrating distributed services over HTTP, including MongoDB ingestion, internal Bosch metadata sources, and a secured PostgreSQL layer.
- Developed a FastAPI-based MCP server on top of the database, enabling brake engineers to query test data using natural language instead of writing SQL or Python.
- Designed and normalized a PostgreSQL schema with validation layers, replacing Excel/CSV/XML-based storage with a more reliable, structured and centralized place for test data.
- Refactored undocumented legacy scripts for automotive test bench data into modular, documented Python packages, improving maintainability and reducing dependency on scattered processing workflows.
Projects
Detection of Freezing of Gait Events in Parkinson’s Disease Using Foot-Worn IMUs
- Built a Freezing of Gait detection pipeline on foot-worn IMU data from Parkinson’s patients, evaluated with patient-wise leave-one-out validation; achieved up to 86.34% event-level F1, showing the system could detect real freezing episodes in noisy daily-life recordings.
- Designed a two-level annotation protocol and temporal undersampling method, reducing redundant non-FoG windows by 45–55% while preserving rare FoG events and improving label quality for a clinically ambiguous task.
- Benchmarked Random Forest, XGBoost, CNN-ResNet18, and CNN-LSTM models, quantifying how each approach balances event sensitivity and reliability under different detection thresholds.
Visual Inspection with Transfer Learning and Domain Adaptation
- Built a deep learning computer vision pipeline for electrical motor coil inspection, addressing domain shift between linear-winding and visually similar but distinct needle-winding production environments.
- Benchmarked Vision Transformers (DINOv2) and CNNs (EfficientNet) for multi-label defect detection, established 87.2% macro-F1 as supervised baseline and retained 82.8% macro-F1 using Unsupervised Domain Adaptation (UDA).
- Demonstrated the viability of bypassing costly manual data annotation, validating a transferable AI inspection framework that can significantly reduce deployment costs for future manufacturing variants.
Leaderboard Standings
Leaderboard Position Pending
Global test scores, peer standing percentiles, and algorithm leaderboard ranks are updated dynamically.
Assessment Highlights
Assessments Not Completed
Coding evaluations, system assessment results, and conceptual score badges will appear here after taking a test.
AI Collaboration Score
AI Collaboration Score Pending
Developer coding behavior, assistant cooperation, and AI pair-programming indicators are evaluated during live coding sessions.
Role Compatibility Profile
Role Compatibility Analysis Pending
Custom matching reports, candidate role compatibility percentiles, and core engineer strength profiles are processed once conceptual code screenings are complete.
Achievements
Achievements Not Earned
Special rewards, developer badges, system recognition certificates, and conceptual milestones will display here.
About Details
Professional Bio
M.Sc in Data Science with background in Physics and hands-on experience at Bosch. I work with messy real-world data, building reliable analysis and ML workflows, and turning results into something engineers, researchers, or business teams can actually use.
Master of Science - Data Science in Data Science
Friedrich-Alexander-Universität Erlangen-Nürnberg (2022 - 2026)
Master of Science - Physics in Physics
The Maharaja Sayajirao University of Baroda (2019 - 2021)
Bachelor of Science - Physics in Physics
Veer Narmad South Gujarat University (2016 - 2019)