Ruchi Mali

A.I. Engineer / Researcher

Exploring the width of A.I.

Connect with me
A little bit about me
Professional Experience
Independent AI Researcher March 2025 - Febraury 2026
  • Developing research papers based on new ideas and novel tecnhiques as well as old ideas as developed during academic years, currently targetting research publications.
Founding AI Engineer June 2024 - March 2025
  • Developed many international client projects, company standalone projects and open source projects.
  • Developed Synthetic dataset generation pipeline (open source) with support for LLMs across all platforms and using novel techniques for synthetic data generation like cot, async request, cache management etc
  • Developed LLM Chat Simulations in alt vibe
  • Developed a Legal LLM, ( US UK and Europe jurisdiction ) with extensive research, data collection and generation and varied finetuining methods as well as developed whole score evaluation system for us/uk client.
  • Worked on GPUS, async requests, cache optimizations, logging, Web To Slack notifications, llm compression techniques kvpress, etc.
  • Worked on Full stack project Notify GPT with web scrapping
  • Created subsystem as in wizardlm research paper
  • Utilized Model Inference Techniques like cot (chain of thought), pvg ( Prover Verifier Games), rto (reinforced token optimization), etc.
  • Conducted Blockchain mining for Agentao and Condense AI (used agent architecture like Swe Agent, Network of multiple LLMS, prompt engineering etc).
  • Developed test cases manual and automatic with pytest.
  • Finetuning LLMs, with experiments, evaluation against benchmarks, used un-censoring techniques like dpo, abliteration, kto, etc.
  • Extensively worked on run pod, modal, AWS remote servers, cursor, windsurf.
A.I. and Data Science Intern December 2023 - April 2024
  • Worked on image to text, voice to text, text to voice, LLMS, yolo, python, flask, APIs Testing, neural networks, OCR, data acquisition, NLP, RAG, Lang chain Framework, Aiding in Model Deployment on Virtual Machine, Optimizing code.
My Independent work
Computer Vision Novel Methods for Handwriting Analysis to detect psychological and behavioural problem and aid the counselling process. 2023
  • Implemented Novel CV Methods for identifying features of Handwriting and Classifying them into groups of psychological, behavioural problems to aid counselling process in Schools. Used Unsupervised learning (K modes, KNN), real data collected from school.
Association Mining implementations 2023
  • Novel method for optimising Apriori Algorithm. o Implemented Apriori and FP Tree on different datasets.
Genetic Algorithm and Ant Colony Optimisation 2022
  • Scratch development of Genetic Algorithm and Ant Colony Optimisation.
Relation on Reading Quotes and Emotions 2022
  • On ground collection of data with questionnaire and finding support for the hypothesis, “Positive Quotes influence positive thoughts”.
Recipe Generator with Deep Learning 2022
  • Conditional Transformer (Hugging Face) to generate Novel Recipes from the Ingredients given, a full end to-end web application for generating recipes either from existing database or creating new ones with text generation. Full stack website.
URL Phishing prediction 2022
  • Predicting if URL is phished or not with Random Forest Classifier and feature extraction methods.
Solar Flare Prediction 2021
  • Predict the possible regions of solar flares with SVM by analysing the rays data of sun.