Shipra
I am a data science student at the Indian Institute of Technology Madras with hands-on experience in AI engineering, quantitative research, software development, and machine learning.
About
I am a data science student at the Indian Institute of Technology Madras with hands-on experience in AI engineering, quantitative research, software development, and machine learning. I enjoy building practical AI systems, including retrieval-augmented generation pipelines, intelligent agents, evaluation harnesses, and high-performance APIs. My work has involved Python development, model evaluation, systematic trading research, natural language processing, and open-source contributions to Keras and related machine learning tooling.
I am especially interested in creating reliable, measurable, and user-focused technology. I have worked on projects involving automated data analysis, virtual teaching assistants, emotion classification, and grounded question-answering systems. Through internships, competitions, scholarships, and open-source work, I have developed a strong foundation in applied machine learning, software engineering, research, and analytical problem-solving.
- 2 years' experience
- Joined September 2026
Industry experience
- Computers & Electronics
- Education
- Financial Services
- Software & Internet
- Professional Services
- Life Sciences
Work experience
Designed and deployed a retrieval-augmented generation answer flow using FastAPI and Supabase, with strict source grounding to reduce hallucination risks in production. Architected schema-strict API contracts using Zod to provide predictable, type-safe JSON payloads. Built an automated evaluation harness and rubric-based testing workspace for benchmarking model outputs against edge cases and token failure modes.
Developed, coded, and backtested systematic trading algorithms using the WorldQuant BRAIN Python API. Optimized hyperparameters and evaluated performance using Sharpe ratio, fitness, turnover, and cross-alpha correlation. Conducted quantitative research on 10 years of market data to identify market inefficiencies, select robust alpha signals, and reduce cross-signal redundancy.
Reviewed and validated AI-generated Python code for correctness, efficiency, and alignment with project specifications. Authored robust Python solutions to improve and diversify datasets for code-generation models. Conducted repeated agent trials and failure analysis, refining task scopes, evaluation logic, edge cases, and anti-hardcoding controls to measure genuine problem-solving and generalization.
Education
Certifications and awards· 9
Hire Shipra today
Post a job and invite Shipra, or contact them directly. No hidden commission.