# Hi, I'm Rishikesh 👋

### 🧠 AI Research Engineer | 🎙️ Voice AI · LLMs · Agentic Systems · ⚡ Efficient Inference

I'm an **AI Engineer & Researcher with ~8.5 years of experience** in Generative AI, speech synthesis, LLM architectures, agentic systems, and AI infrastructure.

I enjoy **implementing research papers from scratch, training and optimizing models, exploring GPU-level inference, and building real-world AI systems.**

- 🚀 **Co-founder & CTO at Dubpro.ai (DeepSync Technologies)** — building speech synthesis, multilingual AI, voice cloning, and generative audio technologies.
- 🎙️ **Former AI Consultant at [MyClone.is](https://www.myclone.is)** (San Francisco) — worked on real-time Voice AI, digital personas, RAG, memory, and agent orchestration.
- 📩 **Email:** [rishikksh20@gmail.com](mailto:rishikksh20@gmail.com)

## 🔬 Research Interests & Engineering

- **LLMs & Architectures:** Transformers, Hybrid/Linear Attention, SSMs, Mamba, model implementations from scratch.
- **Training & Post-training:** SFT, RLHF, RLVR, GRPO, model distillation, and alignment.
- **Efficient Inference:** Triton, CUDA, Flash Attention, KV caching, quantization, GPU memory and latency optimization.
- **Voice & Generative Audio:** TTS, voice cloning, vocoders, neural codecs, diffusion, flow matching, real-time speech systems.
- **Agentic AI:** Coding agents, MCP, multi-agent orchestration, RAG, persistent memory, tool calling, and sandboxing.

## 🚀 Featured AI Agent Projects

- **[MyClone](https://github.com/myclone-dev/myclone)** — Real-time Voice AI digital persona platform with RAG, persistent memory, voice cloning, and persona-aware conversations. `LiveKit` `LlamaIndex` `FastAPI` `pgvector`

- **[Nexus Code Agent](https://github.com/rishikksh20/nexus-code-agent)** — CLI-first AI coding agent featuring MCP, sub-agents, skills, tools, session memory, context management, and sandboxed execution. `AI Agents` `MCP` `Multi-Agent`

- **[Voice AI Agent](https://github.com/rishikksh20/voice-ai-agent-basic)** — Real-time conversational AI with inbound/outbound Twilio calls, browser WebRTC, and text chat. `Pipecat` `Twilio` `STT` `LLM` `TTS`

## 🧠 LLM & Architecture Implementations

Minimal **PyTorch implementations of modern model architectures**, focusing on attention, SSMs, caching, decoding, and inference efficiency.

- **[LFM2.5 2.6B](https://github.com/rishikksh20/lfm25-pytorch)** — Hybrid convolution/attention LLM with Triton inference kernels.
- **[Mamba-3](https://github.com/rishikksh20/mamba3-pytorch)** — Selective State Space Model implementation.
- **[Qwen3.5 0.8B](https://github.com/rishikksh20/qwen3-5-playground)** — Hybrid GatedDeltaNet + GQA architecture.
- **[Gemma 3 270M](https://github.com/rishikksh20/gemma3-270m-playground)** — Lightweight LLM implementation and exploration.

## 🔊 Speech & Generative Audio Research

Selected open-source implementations from my work in speech synthesis, neural audio, and generative modeling:

- **[Voxtral Codec](https://github.com/rishikksh20/voxtral-codec-pytoch)** — Neural audio codec with semantic/acoustic quantization.
- **[NU-Wave](https://github.com/rishikksh20/NU-Wave-pytorch)** — Diffusion-based neural audio upsampling.
- **[iSTFTNet](https://github.com/rishikksh20/iSTFTNet-pytorch)** — Lightweight neural vocoder.
- **[VocGAN](https://github.com/rishikksh20/VocGAN)** — Real-time GAN-based neural vocoder.
- **[AcademiCodec](https://github.com/yangdongchao/AcademiCodec)** — Neural audio codec research.

**Other paper implementations:** [ViViT](https://github.com/rishikksh20/ViViT-pytorch) · [CvT](https://github.com/rishikksh20/convolution-vision-transformers) · [LocalViT](https://github.com/rishikksh20/LocalViT-pytorch) · [FNet](https://github.com/rishikksh20/FNet-pytorch) · [CoaT](https://github.com/rishikksh20/CoaT-pytorch) · [SiT](https://github.com/rishikksh20/SiT-pytorch)

## 🛠️ Tech Stack

**AI & Research:** PyTorch · Transformers · SSMs · RL/RLHF · Diffusion · TTS

**Inference:** Triton · CUDA · Flash Attention · vLLM · KV Cache · Quantization

**Agents & Voice:** MCP · RAG · LlamaIndex · LiveKit · Pipecat · Twilio

**Backend & Infra:** Python · FastAPI · PostgreSQL · Redis · Docker · AWS · GCP

---

### 🤝 Let's Connect

Open to **AI research collaborations, challenging engineering problems, and ambitious AI projects.**

🐙 [GitHub](https://github.com/rishikksh20) · 𝕏 [@ai_rishikesh](https://twitter.com/ai_rishikesh) · 📩 [Email](mailto:rishikksh20@gmail.com)