MS Computer Science student at Northeastern University with 3+ years of industry experience at Accenture as an Application Development Analyst — building enterprise SAP systems and ML-powered applications that operate at scale.
I bridge two worlds: enterprise backend engineering (SAP ABAP, S/4HANA, ETL pipelines, REST APIs) and applied machine learning (computer vision, NLP, semantic retrieval). I don't just train models — I ship full production systems end to end.
- 🏢 Built & maintained enterprise SAP ecosystems processing millions of records daily at Accenture
- 🔬 Researching multimodal clinical AI for skin cancer diagnosis at Northeastern
- 🤖 Exploring SAP Joule (SAP's generative AI copilot) for intelligent ERP automation
- 📍 Boston, MA
Delivered enterprise SAP solutions across SAP ECC, S/4HANA, SAP HANA, and SAP Fiori for large-scale clients.
SAP ABAP & Backend Engineering
- Engineered an ETL data migration pipeline transferring 2 million records from SAP ECC to S/4HANA using standard BAPIs, achieving 100% data accuracy through rigorous schema validation
- Developed automated distributed batch processing with XML-based data transmission, fault-tolerant retry logic, and schema validation — reducing turnaround from 10.4 → 6.0 hrs and cutting manual effort by 36% (~58 hrs/month)
- Diagnosed and resolved a critical production defect, restoring 500+ daily automated workflows and reducing downtime by 40%
- Resolved ~180 monthly production incidents across distributed enterprise systems, maintaining 98% SLA compliance
SAP Fiori · OData · SAP Gateway · SAP Connectors
- Integrated SAP Fiori web frontend with backend services via OData REST APIs (SAP Gateway), following SAP architecture best practices
- Led code reviews and implemented automated unit tests, improving Fiori page load time by 20% and enforcing regression coverage
- Designed and maintained SAP connector interfaces for cross-system data exchange, enabling reliable XML/JSON data transmission across distributed SAP landscapes
SAP HANA
- Worked on SAP HANA as the in-memory database layer backing S/4HANA workloads — optimizing data models, CDS views, and query performance for high-throughput enterprise transactions
SAP Joule (Generative AI for ERP)
- Hands-on exploration of SAP Joule, SAP's embedded generative AI copilot, for intelligent workflow automation, natural-language ERP queries, and AI-assisted incident triage within the SAP Business Technology Platform (BTP)
- Built RESTful microservices and APIs in C# (ASP.NET) with transaction management for the LMS Markers Pro platform
- Optimized API performance by 40% via async/await patterns and implemented data validation for improved reliability
6-class clinical skin lesion classifier fusing visual and textual patient data with Explainability.
- Built on PAD-UFES-20 using MedSigLIP (vision) + ClinicalBERT (text) with cross-attention fusion
- Benchmarked across 10 text encoders; includes fairness evaluation and XAI (Explainable AI) visualizations
- Deployed as a Streamlit interactive demo at derm-platform.me
- Stack: PyTorch · scikit-learn · OpenCV 4 · HuggingFace · Streamlit · XAI
High-performance dense retrieval pipeline with LLM embeddings and approximate nearest-neighbor indexing.
- Used Sentence-Transformers for LLM embeddings + FAISS (IVF-Flat & IVFPQ) for scalable ANN search
- Achieved 95% recall@10, 80% smaller index, and 5× faster queries vs. exact search
- Covers NLP, Generative AI, and Recommendations/Ranking use cases
- Stack: Python · FAISS · Sentence-Transformers · HuggingFace · pytest
Neural signal decoding system for motor-imagery tasks, built for a BCI competition track.
- EEG signal processing and classification pipeline using temporal convolution networks
- Frequency-domain feature extraction for motor-imagery classification
- Stack: Python · MNE · PyTorch · NumPy
End-to-end ML web app: OpenCV face detection → wavelet features → scikit-learn classifier → live deployment.
- OpenCV face detection + PyWavelets feature extraction + scikit-learn SVM classifier
- Flask REST API backend with JavaScript/Bootstrap frontend; deployed live on Render
- Stack: Python · Flask · scikit-learn · OpenCV · NumPy · PyWavelets · JavaScript · Bootstrap
Systematic benchmark of CNN and ViT architectures on visual recognition tasks.
- Compared ResNet, EfficientNet, and Vision Transformer (ViT) on recognition benchmarks
- Custom training loops, data augmentation, and ablation studies
- Stack: PyTorch · torchvision · Matplotlib
Real-time camera calibration pipeline with 3D AR object overlay.
- Full camera calibration (intrinsics, extrinsics, distortion) + pose estimation for AR overlays
- Stack: Python · OpenCV · NumPy
Real-time lost-and-found platform with geolocation, image upload, and category filtering.
- REST API backend with JWT auth, image upload, and geolocation filtering; React frontend
- Stack: Node.js · Express · MongoDB · React · Render
Restaurant ordering and menu management application with full CRUD operations.
- Stack: Node.js · Express · PostgreSQL · React · Render
ML / AI
SAP Ecosystem
Languages
Web & Cloud
| Badge | Certification | ID |
|---|---|---|
 |
Google Cloud Certified — Professional Cloud Architect | 100076 |
 |
Google Cloud Certified — Associate Cloud Engineer | 139218 |
3+ years of enterprise engineering · MS Computer Science @ Northeastern · Google Cloud Certified
Open to full-time roles in ML Engineering, Software Engineering, and SAP Development.