Experience
From freelance Python to shipping AI products for clients.
Aug 2024 – now● Current
Backend Engineer · Newtuple Technologies
Pune
A GenAI services company that ships complete, deployed software for clients. I joined when the team was about ten people, so the work has spanned services, auth, infrastructure, AI features and engineering practice.
- Took a care-home platform for a UK startup from pilot to production: owned the architecture, cloud and security, and built AI notes and an email-driven referrals portal.
- Built a citation-backed RAG chatbot with real-time text and speech for a US national website.
- Extended a text-to-SQL agent with streaming, tracing, load balancing and an evaluation suite, and cut its base prompt from ~20k to ~7k tokens.
- Interview candidates at every seniority level, review PRs and support the junior developers day to day.
Case study 01: Care-home management platform →Case study 03: Citation-backed RAG chatbot →Case study 02: SQL chat agent →
Oct 2022 – Aug 2024
Python Developer · Freelance
Remote
Two years of building AI, automation and data-processing tools for clients, start to finish.
- A Windows app that lets non-technical users train and deploy custom YOLO object-detection models.
- A transcription pipeline built on OpenAI Whisper.
- An OCR and table-extraction tool that turns paper documents into structured text.
- A product-comparison scraper with Selenium, BeautifulSoup, Flask and Redis.
2019 – 2023
B.Tech · Government College of Engineering
Aurangabad
Where the machine-learning side started: computer vision, ML pipelines and an LLM chatbot, built as student projects.
- Explainable industrial defect detection ↗
Classifies product defects and locates the defective area in an image without manual labelling of the defect region, using a VGG16-based architecture.
- APS fault detection for heavy-duty vehicles ↗
An end-to-end ML pipeline covering ingestion, validation, transformation and model deployment.
- AI-powered WhatsApp chatbot ↗
An LLM connected to WhatsApp through the Twilio API, with custom tools.
- Explainable industrial defect detection ↗
Contributions and guidance
Helping the people and projects around me move faster.
- Guiding developers. I support four to five developers, mostly juniors, day to day on architecture decisions, git workflow and deployments.
- Code review. I review 20–30 PRs a week on my own project, plus three to five a week on others, keeping quality consistent across the codebase.
- Interviewing. I interview candidates across all seniority levels, from freshers to senior engineers, and have conducted more than 50 technical interviews.