yashwadgave!

Yash@MacBook — Terminal

██╗   ██╗ █████╗ ███████╗██╗  ██╗
╚██╗ ██╔╝██╔══██╗██╔════╝██║  ██║
 ╚████╔╝ ███████║███████╗███████║
  ╚██╔╝  ██╔══██║╚════██║██╔══██║
   ██║   ██║  ██║███████║██║  ██║
   ╚═╝   ╚═╝  ╚═╝╚══════╝╚═╝  ╚═╝

██╗    ██╗ █████╗ ██████╗  ██████╗  █████╗ ██╗   ██╗███████╗
██║    ██║██╔══██╗██╔══██╗██╔════╝ ██╔══██╗██║   ██║██╔════╝
██║ █╗ ██║███████║██║  ██║██║  ███╗███████║██║   ██║█████╗  
██║███╗██║██╔══██║██║  ██║██║   ██║██╔══██║╚██╗ ██╔╝██╔══╝  
╚███╔███╔╝██║  ██║██████╔╝╚██████╔╝██║  ██║ ╚████╔╝ ███████╗
 ╚══╝╚══╝ ╚╝  ╚╝╚═════╝  ╚═════╝ ╚╝  ╚╝  ╚═══╝  ╚══════╝

I love working through messy systems/workflows and stick with them until they work in prod!

Reach out to me on Twitter and see my work on GitHub.

loading work_history...

✱AI Engineer @ Square Yards

Nov 2025 — Present • Mumbai

-Improved the voice-agents post-call harness to identify accurate leads and reduce false positives, taking the combined system accuracy to 94.33% and helping RMs focus on the right leads.

-Built custom MCP servers that power 4+ internal AI apps along with ChatGPT app and Claude connectors, contributing an additional 2.7% lead acquisition channel.

-Revamped the existing platform into a ChatGPT-like real estate AI platform with persistent memory, dynamic user context, and custom harnesses to route leads to RMs and Sales, increasing lead conversion from ~6.4% to 11%.

✱Founder/CTO @ Tisac AI Technologies

May 2023 — Oct 2024 • Hybrid

-Built enterprise AI tools for data labeling, model training/deployment, and Slack-based customer support workflows.

✱Data Science Intern @ TruBoard Partners

Jan 2022 — Jun 2022 • Mumbai

-Built solar power prediction models and loan-document extraction workflows using Python, scikit-learn, and Azure OCR.

loading selected_projects...

Transformer Model (Python)(AI/Models)

Built a small decoder-only transformer from scratch to learn all the nuances of model building, reached an exhaustive accuracy of 99.78% for arithmetic operations, also post-trained it using SFT and explored RL using GRPO for improving existing capability. Concluded with data/representation mattered more in pre-train than RL unless the model is already good.

Agent Builder(AI Agents/DevTools)

Built a full-stack local AI agent builder in Go + Next.js with SQLite-backed agent specs, a shared Go run engine, SSE chat streaming, a tool registry with approval gates, edit-agent chats, source tracking, version snapshots, and OpenAI-powered custom agents.

Deep Research Agent(AI Agents/Research)

Built a from-scratch TypeScript deep-research agent directly on the OpenAI raw HTTP API, without the Agents or Assistants SDK, to own the while-loop harness end to end. The LLM supplies intent while deterministic code handles Zod argument validation, tool dispatch, recursion, persistence, isolated recursive sub-agents, Exa-powered web search and reading, and a live terminal UI that streams cited answers.

Open Office XML Edit Infra(AI/Document Infra)

Built infrastructure for reliable agentic Office document editing by converting PPTX/OOXML files into canonical graph representations and applying deterministic document mutations. Solves the reliability gap in automated white-collar document workflows through structure-aware transformations, layout-preserving edits, and render validation for long-tail document operations where current AI fails (last 10% editing) to perform actions accurately.

AI Agent Runtime Audit and Governance SDK(AI/Compliance)

Built a Python SDK for auditing and governing AI agent runtimes by capturing execution events, enforcing policies and guardrails, redacting sensitive data, and generating tamper-evident evidence packs with hash-linked integrity for compliance review and operational oversight. Open-sourced it with Apache-2.0 license.

Memory-engine(AI/Memory)

Built an experimental memory framework for LLM agents to store, retrieve, decay, and reason over long-term memories using SQLite-backed local storage. Added cleanup logic that can rewrite, invalidate, and remove stale memories to keep context useful and redundancy-free.

Agentic Coding Tool(DevTools/Agents)

Built a Python coding assistant with sandboxed shell execution, file tools, streaming OpenAI responses, persistent sessions, and path validation for safer agent workflows.

Operating System Language Model(AI/Systems)

Built a custom OS-focused language model in PyTorch for command-aware assistance, with a guarded training pipeline, tokenizer, CLI, and FastAPI runtime.

compiling certifications...

✓[CREDENTIAL] Analyze Datasets and Train ML Models using AutoMLVERIFIED

source:Coursera

verify:[BZ75BTL4VD9Z]

✓[CREDENTIAL] Build, Train, and Deploy ML Pipelines using BERTVERIFIED

source:Coursera

verify:[3STZG4V889HW]

✓[CREDENTIAL] Optimize ML Models and Deploy Human-in-the-Loop PipelinesVERIFIED

source:Coursera

verify:[LYMJFK9LAKMG]

✓[CREDENTIAL] Practical Data Science on the AWS CloudVERIFIED

source:Coursera