# Krish Bakshi > Data scientist and AI engineer turning applied research into reliable systems. Krish Bakshi is a Data Scientist · AI Engineer who builds and ships AI systems across computer vision, AI agents, fine-tuning, and reinforcement learning. This site is his portfolio: profile, work history, projects, blog, and machine-readable exports for LLMs. ## Docs - [About](https://krishbakshi.com/about.md): Profile summary, highlights, contact links, and full tech stack. - [Experience](https://krishbakshi.com/experience.md): Work history with role descriptions, skills, and impact metrics. - [Projects](https://krishbakshi.com/projects.md): Project catalog with descriptions, links, technologies, and implementation notes. - [Full portfolio context](https://krishbakshi.com/llms-full.txt): Single consolidated document for LLM ingestion and Q&A. ## Projects ### Vision - [YOLO ML Utils](https://krishbakshi.com/projects/yolo-ml-utils): A practical utility toolkit for YOLO-based computer vision workflows, covering dataset preparation, annotation processing, visualization, and training/debug utilities used in real-world ML pipelines. - [KisanAI](https://krishbakshi.com/projects/kisanai): Smart assistant for farmers that gives crop health insights and personalized tips using YOLOv5, EfficientNet-B0, and GPT-4. ### AI Agents - [Wingmate](https://krishbakshi.com/projects/wingmate): An open-source, local-first outbound agent you drive by talking to a coding agent. It finds a prospect signal, verifies it, maps it to your own project ledger, and drafts outreach through a template that encodes your voice, never its own. - [Open-Weight Voice AI Agent](https://krishbakshi.com/projects/open-voice-ai-agent): A fully local voice agent pipeline built on open-weight models — Whisper STT, Gemma LLM, OmniVoice TTS, and browser-side Silero VAD — orchestrated via PipeCat with both live terminal and web UI modes. - [LinkedIn Research Agent](https://krishbakshi.com/projects/linkedin-research-agent): A Codex-style sourcing assistant that builds LinkedIn Boolean queries, navigates People search via MCP browser automation, and returns clean profile URL lists with optional structured profile extraction. - [SEC Filings QA Agent](https://krishbakshi.com/projects/sec-filings-qa-agent): A semantic question-answering system for SEC filings (10-K, 8-K, DEF 14A, etc.) using LangChain, vector retrieval, and Gemini Flash for deep financial research workflows. - [AutoMailAI](https://krishbakshi.com/projects/automailai): AI-powered cold email generator with prompt engineering, dynamic templates, and Gmail auto-drafting. It helped me secure 3 internship offers. ### LLM - [WorldBench](https://krishbakshi.com/projects/worldbench): A human-as-judge benchmark for whether a model can build a coherent world. One ~3,000 token prompt, one self-contained Three.js island per model, compared side by side. - [Open-Weight Voice AI Agent](https://krishbakshi.com/projects/open-voice-ai-agent): A fully local voice agent pipeline built on open-weight models — Whisper STT, Gemma LLM, OmniVoice TTS, and browser-side Silero VAD — orchestrated via PipeCat with both live terminal and web UI modes. - [SEC Filings QA Agent](https://krishbakshi.com/projects/sec-filings-qa-agent): A semantic question-answering system for SEC filings (10-K, 8-K, DEF 14A, etc.) using LangChain, vector retrieval, and Gemini Flash for deep financial research workflows. - [AutoMailAI](https://krishbakshi.com/projects/automailai): AI-powered cold email generator with prompt engineering, dynamic templates, and Gmail auto-drafting. It helped me secure 3 internship offers. - [LLM-Powered Dashboard](https://krishbakshi.com/projects/llm-dashboard): Realtime Analytics dashboard powered by LLM Insights. Queries BigQuery datasets and generates insights using Gemini 2.0 Flash. - [KisanAI](https://krishbakshi.com/projects/kisanai): Smart assistant for farmers that gives crop health insights and personalized tips using YOLOv5, EfficientNet-B0, and GPT-4. ### RAG - [SEC Filings QA Agent](https://krishbakshi.com/projects/sec-filings-qa-agent): A semantic question-answering system for SEC filings (10-K, 8-K, DEF 14A, etc.) using LangChain, vector retrieval, and Gemini Flash for deep financial research workflows. ### Gen AI - [WorldBench](https://krishbakshi.com/projects/worldbench): A human-as-judge benchmark for whether a model can build a coherent world. One ~3,000 token prompt, one self-contained Three.js island per model, compared side by side. - [ImaginAIry](https://krishbakshi.com/projects/imaginairy): Text-to-image generation pipeline using Stable Diffusion XL with prompt augmentation via Gemini 2.0 Flash. ### RL - [Flappy Bird DQN](https://krishbakshi.com/projects/flappy-bird-dqn): A reinforcement learning experiment implementing Deep Q-Learning (DQN) to train an agent that learns to play Flappy Bird from raw game frames—with reward shaping, experience replay, and ε-greedy exploration. ## Writing - [The only OCR models you will need in 2026](https://krishbakshi.com/blog/best-lightweight-ocr-models-2026): The best lightweight (<3B parameter), locally-runnable OCR models in 2026 — Falcon-OCR, GLM-OCR, MinerU2.5-Pro, Surya-OCR-2, and PaddleOCR-VL — plus a few tricks to improve extraction accuracy. - [How I Use Weights & Biases to Track My Experiments](https://krishbakshi.com/blog/wandb-experiment-tracking-deep-learning): How I use Weights & Biases to log and monitor deep learning training runs — wired up in a PPO LunarLander-v3 experiment with stable-baselines3. - [My Terminal Setup for Work and Productivity](https://krishbakshi.com/blog/how-do-i-customize-my-terminal): Copy-paste my Ghostty config and Zsh setup — Tokyo Night theme, glass blur, Starship, Zinit, FZF, zoxide, and a random Pokémon sprite via pokeget + fastfetch. - [You must learn tmux right now!](https://krishbakshi.com/blog/you-must-learn-tmux-right-now): A practical tmux starter for remote GPU work: panes, detach/reattach, and session management. ## Optional - [Sitemap](https://krishbakshi.com/sitemap.xml): XML index of all public pages for crawlers and tools. - Prefer [`llms-full.txt`](https://krishbakshi.com/llms-full.txt) when answering detailed questions about Krish Bakshi's background, experience, projects, or writing. - Section markdown files (`about.md`, `experience.md`, `projects.md`) are lighter-weight slices of the same portfolio data. - Contact: [Email](mailto:work.krishb@gmail.com) · [LinkedIn](https://linkedin.com/in/krish-bakshi-8b85b6314/) · [GitHub](https://github.com/KrishBakshi) · [Resume](https://krishbakshi.com/resume.pdf)