I build cloud data systems, ML-enabled workflows, and backend platforms that turn messy business data into reliable products.
Focused on
A place for technical essays, implementation notes, and lessons learned while building AI systems, data pipelines, and backend products.
A build log on the browser-first tool I made to turn weak source narration into a cleaner AI voiceover pipeline using Groq transcription, Orpheus speech, and client-side video rendering.
Read post
A build log on two very different systems that share the same goal: turn dense source material into useful answers without letting the model drift away from the truth.
Read post
A detailed build log on the local-first AI pipeline I made to tailor resumes against job descriptions while preserving evidence, structure, and ATS-friendly LaTeX output.
Read post
A reflection on moving fast with AI-assisted development while using CI/CD guardrails to keep code, data pipelines, and production systems trustworthy.
Read post
A project story on BLIP-2, LLaVA, VizWiz, PyTorch, Hugging Face, React, and Flask for real-time auditory feedback.
Read post
A USC data modeling project on internal PostgreSQL code changes, join operator heuristics, and query planner behavior.
Read post
A project story on ResNet-50, Grad-CAM saliency maps, and clinician-facing real-time model interpretation.
Read post
A differential testing framework for Wasmtime, Wasmer, and V8 that turns runtime behavior into analyzable artifacts.
Read postSelected work from my AI, data engineering, analytics, backend, and systems resumes. The common thread: take a hard data problem, design the workflow around trust and scale, and make the result usable.
Static GitHub Pages tool that extracts speech from uploaded video, uses a user-provided Groq key for transcription and Orpheus voice generation, then renders a downloadable video locally.
Built an AI-powered resume tailoring pipeline that reads a resume plus job description, rewrites only the relevant evidence-backed bullets, and regenerates a polished PDF from the original LaTeX template.
Built a real-time assistive AI system combining Faster R-CNN object detection, BLIP-2 image captioning, OCR, and NLP-driven auditory feedback.
Designed a clinician-facing gastric cancer screening pipeline using ResNet-50 and Grad-CAM saliency maps for interpretable real-time predictions.
Automated clinical-trial ingestion, validation, deduplication, and fuzzy matching workflows for structured healthcare data using Python, SQL, and AWS services.
Architected a distributed OLAP-style query engine using Java and Apache Pekko typed actors for multi-cluster graph analytics workloads.
Engineered a three-class dermoscopy classifier using EfficientNet, Grad-CAM overlays, SMOTE, class weighting, and augmentation for lesion interpretation.
Optimized PostgreSQL B-Tree index prefetching and scan behavior for OLAP benchmark queries using C, Docker, GDB, and JOB/IMDB workloads.
Migrated RDF/SPARQL graph logic to Neo4j for customer microsegmentation, KOL identification, and graph-based feature analysis.
Used PySpark, AWS, EDA, statistical analysis, and Tableau to identify outliers and propose top diversity-rich clinical sites for trial planning.
Led a Flask-based predictive analytics app with preprocessing and feature engineering to forecast flight price fluctuations.
Built and evaluated a logistic regression model for fraudulent transaction detection with data cleaning and feature engineering.
Converted video to images, isolated human objects with TensorFlow, measured distances, and highlighted social distancing violations.
Developed Angular and Flask API flows for Excel uploads, using SQL and Python microservices to improve data-heavy planner workflows.
Deployed a Java-based AWS REST API to generate keyword weightings from submitted text and tested reliability with Postman on 500+ samples.
Led a real-time geolocation system with Firebase portals for drivers and traffic police, automating proximity alerts for emergency response.
GitHubA contribution-style systems map for people who want to inspect how I build across AI data platforms, backend systems, DevOps, and applied research. Hover or tap the grid, then follow the signal in the console.
Hydrated from a scheduled GitHub Actions export of real contribution and repository language data.
Sajid's clear communication and strategic thinking foster a collaborative and productive environment, enhancing cross-functional cooperation.
Palak Dubey, Reporting Manager at ZS
His quick learning and problem-solving skills ensure timely project deliveries and consistent outperformance.
Vishal Patil, Reporting Manager at DeeDee Labs
Sajid's continuous pursuit of knowledge and strong command of Python and Databricks make him a valuable asset to any team.
Dhanshri Jangam, Data Scientist
Alongside engineering, I mentor students and early-career professionals on resumes, interviews, networking, and career direction through focused 1:1 sessions.
Hello, you are viewing Sajid's AI and Data Engineering portfolio. Ask about his technical expertise, projects, mentorship, or collaboration options. You can ask in English or Spanish.
