Doyel Mishra.
Building the pipeline underneath the product — where machine learning meets shipped software.
That's the question that keeps pulling me toward AI/ML: not the model in isolation, but the pipeline it needs to survive contact with real data.
Daily DSA practice, open‑source contributions via GirlScript Summer of Code, and a structured AI/ML internship right now — closing the gap between a model and a product.
Two internships. One in progress.
- Structured, curriculum‑based internship spanning AI, Machine Learning, and Generative AI fundamentals.
- Practicing prompt engineering alongside Agile, the SDLC, and Git/GitHub workflows.
- Building backend fluency to connect ML work to deployable applications.
- Designed and optimized 20+ prompts for AI automation workflows, improving output relevance by roughly 30%.
- Engineered 3 AI‑powered mini tools via multi‑step prompt chaining, cutting task time by roughly 25%.
- Built automation pipelines with Zapier and n8n; developed AI music‑ and diagram‑generation tools.
Three problems, shipped.
An AI resume‑analysis tool built on FastAPI and sentence‑transformers. It computes a JD match score and surfaces skill gaps — with an async architecture behind it.
View on GitHub ↗A Streamlit dashboard combining Prophet and polynomial regression — configurable 30–365 day forecasts, KPI tracking, rolling averages, and anomaly alerts.
View on GitHub ↗An NLP pipeline classifying open‑ended employee feedback through text preprocessing and sentiment modeling — turning free text into HR‑actionable signal.
View on GitHub ↗A working system, not a checklist.
hover to trace connections
Hover a mark to read it.
ISRO SDSC SHAR
Selected for an internship at Satish Dhawan Space Centre — India's premier space launch facility.
Open to internships, collaborations, and interesting conversations. Currently available for hire.