Skip to content
View Devesh-Maheshwari's full-sized avatar
6 followers · 11 following

Achievements

Achievement: Pull Shark

Achievements

Achievement: Pull Shark

Highlights

  • Pro

Block or report Devesh-Maheshwari

Block user

Prevent this user from interacting with your repositories and sending you notifications. Learn more about blocking users.

You must be logged in to block users.

Maximum 250 characters. Please don’t include any personal information such as legal names or email addresses. Markdown is supported. This note will only be visible to you.
Report abuse

Contact GitHub support about this user’s behavior. Learn more about reporting abuse.

Report abuse
Devesh-Maheshwari/README.md

Hey, I'm Devesh Maheshwari 👋

Engineering AI systems with a spinner's mindset — patience, precision, and reading the game two moves ahead. 🏏

profile views

🧑‍💻 About Me

  • 🎓 MSCS @ University of Wisconsin–Madison (2026) · B.Tech EEE @ IIT (BHU) Varanasi
  • 🔬 Published at ICLR 2026 Workshop — neurosymbolic reasoning with LLMs
  • 💼 4.5+ years of industry ML experience across Auxia, Spring Financial, HiLabs, Publicis Sapient
  • 🤖 Obsessed with: Agentic Systems, LLM post-training (PPO/DPO), RAG pipelines, Verifiable AI
  • 🏗️ Love to scale my learning — from fine-tuning LLaMA/Mistral to building multi-agent orchestration with LangGraph & Google ADK
  • 🏓🏏 Off the keyboard: smashing table tennis shots & spinning cricket balls
  • 📍 Madison, WI · Open to Work — MLE · AI Engineer · Applied Scientist · Data Scientist · SWE

🚀 Featured Projects

Project Description Tech
verifiable-rl-coder Verifiable reinforcement learning for code generation — reward-shaped RL training Python · PyTorch
ClaimBench Benchmark suite for evaluating LLM-based claim verification at scale Python · HuggingFace
llmpackage Clean toolkit for fine-tuning & few-shot learning with LLaMA and Mistral variants Python · PEFT
KV-store High-performance in-memory KV store — epoll/kqueue, sharded concurrency, LRU + TTL C
CS-739-KVStore Distributed in-memory KV store (UW-Madison systems course) C++
CafeAI Conversational coffee shop ordering app — multi-parser LLM backend (GPT-4o, Claude), React Native + Node.js TypeScript · React Native · Node.js

🛠️ Tech Stack

Languages Python C C++ SQL Bash

ML / AI / LLMs PyTorch HuggingFace LangChain OpenAI Mistral

Infra & Tools Linux Docker Git FastAPI AWS

💼 Experience Highlights

  • Teaching Assistant — UW-Madison · Data Science Programming II (Jan 2026 – Present)
  • Data Scientist — Auxia · RL & personalization systems (Jan 2025 – Aug 2025)
  • Data Scientist — Spring Financial · ML for personal loan decisioning (Aug 2024 – Jan 2025)
  • Data Scientist II — HiLabs · NLP / GenAI for healthcare data (Apr 2023 – Aug 2024)
  • Senior Data Scientist — Publicis Sapient · GenAI & chatbot evaluation frameworks (Jun 2021 – Apr 2023)

📫 Let's Connect

I'm actively looking for MLE · AI/ML Engineer · Applied Scientist · Data Scientist · SWE roles (full-time, 2026). Drop me a line at deveshmaheshwari25@gmail.com or connect on LinkedIn 🚀

Pinned Loading

  1. CS-739-KVStore CS-739-KVStore Public

    P1- in-memory KV store in C++

    Makefile

  2. Logic-LLM Logic-LLM Public

    Forked from teacherpeterpan/Logic-LLM

    The project page for "LOGIC-LM: Empowering Large Language Models with Symbolic Solvers for Faithful Logical Reasoning"

    C

  3. Image_Encryption Image_Encryption Public

    Enhancing Image Encryption and Network Security Through Metaheuristic Optimization

    Python

  4. llmpackage llmpackage Public

    Customized package to ease the finetuning and few shot learning of llama and mistral variants

    Python

  5. KV-store KV-store Public

    High-performance in-memory key-value store in C with epoll/kqueue, sharded concurrency, LRU eviction, and TTL expiry

    C

  6. verifiable-rl-coder verifiable-rl-coder Public

    Python