AI & Software Engineer | Specializing in Computer Vision & Scalable Automation B.E. Computer Engineering (AI/ML) + Honors in Data Science
I bridge the gap between Machine Learning research and production-grade software. I focus on building high-throughput pipelines, automating complex workflows, and ensuring system observability.
| Category | Tools |
|---|---|
| DevOps | Docker, Jenkins, CI/CD, Agile methods, Linux, Git |
| Core Engineering | System Design, SDLC, Data Structures & Algorithms |
| Cloud Computing | AWS (EC2, S3, EB), IBM Cloud |
| AI & Machine Learning | TensorFlow, Scikit-Learn, CNNs, RNNs, Computer Vision, NLP, SigLIP, FAISS |
| Data Analytics | Matplotlib, Seaborn, Pandas, Numpy |
| Languages | Python (Advanced), C, SQL, Dart, Java |
| Mobile & Frontend | Flutter , Html, Css |
- Challenge : Tackling misleading "healthy" marketing claims and decoding complex ingredient lists filled with hidden sugars, harmful additives (E-numbers), and refined flours / oils traps.
- Solution : Engineered Food Transparency system using FastAPI and IBM Watsonx.ai, and implemented LLM-based semantic ingredient analysis.
- Engineering Milestone : Designed a modular Service-Oriented Architecture integrating WatsonX AI and Open Food Facts API, enabling real-time product audits with AI-driven consumer protection guardrails.
- Challenge : Eliminating dependence on cloud-based platforms like Google Photos for semantic video search, where accuracy might be limited and user data is externalized.
- Solution : Engineered a cross-modal retrieval pipeline using SigLIP and FAISS to map natural language queries to high-dimensional frame embeddings.
- Engineering Milestone : Optimized inference to process 100 minutes of video in 21 seconds on standard T4 cloud GPU. Without GPU, on regular phone CPUs it takes around 1-2 minutes for 5 minutes of video file.
- The Challenge: The marketing team manually sent promotional messages on WhatsApp. They wanted automation without purchasing expensive Business APIs or relying on third-party services. Also, they needed customer database management application.
- Impact: Designed and Deployed Automation and database management app. It is operational 2 retail outlets ("Cake of the Day, Ratnagiri").
- Value: Transformed a manual process into a streamlined, automated system, cutting operational overhead and establishing a sustainable, self-managed marketing infrastructure.
- Problem Solving: 200+ LeetCode problems solved (my favorites: Heap, Graph Trversals).
- Certifications: Google IT Automation & Data Analytics. + more than 40 other certeficates. (Too long list to mention here.)
"Turning code into real-world impact, one commit at a time."