Prasad Khake
Full-stack AI engineer · On-device & private LLM deployment
I make LLMs run well on hardware you control — on-device, on-prem, at the edge — and build the products around Most AI ships to someone else's cloud.
More
But plenty of teams can't or won't send their data there — healthcare, finance, legal, or anyone with privacy, cost, offline, or latency constraints. Getting a capable model running privately, inside a real hardware budget, and making it actually usable — that's the work I Lately I've been deep in the on-device stack: contributing fixes to Apple's MLX (the framework for running LLMs on Macs) and publishing honest benchmarks of what really runs on consumer hardware — at com and in my On Device That sits on a decade of shipping: full-stack AI systems (Python data/LLM pipelines, React + TypeScript apps, a HIPAA-compliant healthcare portal), GTM and revenue automation, and an open-source e-paper hardware platform ink) shipped to 20+ countries. The hardware background matters more than it sounds — getting models onto constrained devices is half software, half knowing the If you're trying to run an LLM privately — on a laptop, on-prem, or on an edge device — and it's too big, too slow, or too unreliable, that's exactly the problem I like. Let's Core: on-device & private LLM deployment · quantization & memory optimization · MLX / Apple Silicon · Python · React/TypeScript · full-stack com · com
GTM Index
- STEALTHJOB
- Independent AI Eng3M
Independent AI Engineer — On-Device & Private LLM
Stealth
On-device and private LLM deployment — helping teams run capable models on hardware they control (laptop, on-prem, edge) instead of someone else's cloud. • Contributing to Apple's MLX (mlx-lm) — e.g. a tokenizer-routing fix that broke Mistral/Devstral output on Apple Silicon (PR #1329). • Built & open-sourced ondevice-bench — honest local-LLM benchmarks (speed, memory, what actually fits) on consumer Macs. • Writing on on-device/private LLM deployment at prasadkhake.com and the On Device newsletter. • Open to consulting & build work: model selection + quantization to fit your hardware, private/offline LLM features, edge deployment.
- PAPERD.INKJOB
- Chief Executive Officer6Y 7M
Chief Executive Officer
paperd.ink
- paperd.ink is an open-source e-paper (E Ink) development board — programmable via WiFi, Bluetooth, and USB-C, in the same family as Arduino and Raspberry Pi. - Sold globally across 20+ countries before pausing production due to supply chain constraints.
- CAREER-9JOB
- AI GTM Eng1Y 8M
AI GTM Engineer
Career-9
- Built and shipped AI-powered GTM automation systems end-to-end — LLM APIs (OpenAI, Claude, Gemini), Make.com, n8n, and custom Python/JS glue code. - Designed webhook-driven pipelines connecting CRMs, enrichment tools, and outbound platforms; handled retries, rate limits, and deduplication in production. - Built AI workflow steps for extraction, classification, summarisation, and routing with output quality validation baked in. - Ran Clay and Apollo enrichment pipelines feeding HubSpot and Salesforce; automated cold email sequencing with LLM personalisation at scale. - Built a bulk landing page generation pipeline — CSV-to-SEO-site at scale, with CRM attribution integrated end-to-end.…
- INAZAJOB
- Mktg Mgr2Y 1M
Marketing Manager
Inaza
- Owned marketing automation architecture for a B2B SaaS company — HubSpot, outbound sequencing, and lead attribution pipelines. - Built API integrations between MarTech platforms and internal data systems; wrote enrichment and ICP segmentation scripts in Python and JavaScript. - Wrote long-form technical and thought leadership content for the blog; owned SEO strategy — keyword research, on-page optimisation, and content planning to drive organic growth. - Managed paid acquisition across LinkedIn, Google, and Meta — full-funnel from awareness to pipeline. Partnered with engineering to maintain reliable integrations with logging, alerting, and failure recovery.
- QUANTFARM PVT.JOB
- Senior Business Consultant1Y 1M
Senior Business Consultant
QuantFarm Pvt. Ltd.
- Led AI, analytics, and automation delivery for Fortune 500 clients across multiple industries — owned requirements, architecture, and delivery end-to-end. - Built dashboards, ETL pipelines, and workflow automations that replaced manual reporting and decision-making processes. - Managed client relationships and translated business requirements into technical specs for the development team.
- Business Consultant1Y 10M
Business Consultant
QuantFarm Pvt. Ltd.
- Delivered data integration and automation solutions for MNCs across finance, healthcare, supply chain, and logistics. - Built self-serve dashboards and automated reporting pipelines for executive stakeholders at enterprise companies. - Managed development team execution — translated business requirements into functional and technical specs across concurrent client projects.
- Data Analyst & Programmer1Y
Data Analyst & Programmer
QuantFarm Pvt. Ltd.
- Built analytics solutions across finance, digital media, healthcare, supply chain, manufacturing, and shipping. - Programmed data pipelines, transformations, and reports — early hands-on work across diverse industry data problems.
- Research Assistant3M
Research Assistant
QuantFarm Pvt. Ltd.
- Carried out market analysis and delivered daily intelligence reports for one the largest agro-commodity trading companies in the world
- MAHARASHTRA INSTITUTE OF TECHNOLOGYSCHOOL
- Bachelor of Eng - BE, Elec…4Y 11M
Bachelor of Engineering - BE, Electrical, Electronics and Communications Engineering
Maharashtra Institute of Technology
- Job: Stealth — Independent AI Engineer — On-Device & Private LLM · 2026-05 – present · 3 mo — On-device and private LLM deployment — helping teams run capable models on hardware they control (laptop, on-prem, edge) instead of someone else's cloud. • Contributing to Apple's MLX (mlx-lm) — e.g. a tokenizer-routing fix that broke Mistral/Devstral output on Apple Silicon (PR #1329). • Built & open-sourced ondevice-bench — honest local-LLM benchmarks (speed, memory, what actually fits) on consumer Macs. • Writing on on-device/private LLM deployment at prasadkhake.com and the On Device newsletter. • Open to consulting & build work: model selection + quantization to fit your hardware, private/offline LLM features, edge deployment.
- Job: paperd.ink — Chief Executive Officer · 2020-01 – present · 6y 7mo — - paperd.ink is an open-source e-paper (E Ink) development board — programmable via WiFi, Bluetooth, and USB-C, in the same family as Arduino and Raspberry Pi. - Sold globally across 20+ countries before pausing production due to supply chain constraints.
- Job: Career-9 — AI GTM Engineer · 2024-09 – 2026-05 · 1y 8mo — - Built and shipped AI-powered GTM automation systems end-to-end — LLM APIs (OpenAI, Claude, Gemini), Make.com, n8n, and custom Python/JS glue code. - Designed webhook-driven pipelines connecting CRMs, enrichment tools, and outbound platforms; handled retries, rate limits, and deduplication in production. - Built AI workflow steps for extraction, classification, summarisation, and routing with output quality validation baked in. - Ran Clay and Apollo enrichment pipelines feeding HubSpot and Salesforce; automated cold email sequencing with LLM personalisation at scale. - Built a bulk landing page generation pipeline — CSV-to-SEO-site at scale, with CRM attribution integrated end-to-end.…
- Job: Inaza — Marketing Manager · 2022-06 – 2024-07 · 2y 1mo — - Owned marketing automation architecture for a B2B SaaS company — HubSpot, outbound sequencing, and lead attribution pipelines. - Built API integrations between MarTech platforms and internal data systems; wrote enrichment and ICP segmentation scripts in Python and JavaScript. - Wrote long-form technical and thought leadership content for the blog; owned SEO strategy — keyword research, on-page optimisation, and content planning to drive organic growth. - Managed paid acquisition across LinkedIn, Google, and Meta — full-funnel from awareness to pipeline. Partnered with engineering to maintain reliable integrations with logging, alerting, and failure recovery.
- Job: QuantFarm Pvt. Ltd. — Senior Business Consultant · 2021-06 – 2022-07 · 1y 1mo — - Led AI, analytics, and automation delivery for Fortune 500 clients across multiple industries — owned requirements, architecture, and delivery end-to-end. - Built dashboards, ETL pipelines, and workflow automations that replaced manual reporting and decision-making processes. - Managed client relationships and translated business requirements into technical specs for the development team.
- Job: QuantFarm Pvt. Ltd. — Business Consultant · 2019-07 – 2021-05 · 1y 10mo — - Delivered data integration and automation solutions for MNCs across finance, healthcare, supply chain, and logistics. - Built self-serve dashboards and automated reporting pipelines for executive stakeholders at enterprise companies. - Managed development team execution — translated business requirements into functional and technical specs across concurrent client projects.
- Job: QuantFarm Pvt. Ltd. — Data Analyst & Programmer · 2018-06 – 2019-06 · 1y — - Built analytics solutions across finance, digital media, healthcare, supply chain, manufacturing, and shipping. - Programmed data pipelines, transformations, and reports — early hands-on work across diverse industry data problems.
- Job: QuantFarm Pvt. Ltd. — Research Assistant · 2018-02 – 2018-05 · 3 mo — - Carried out market analysis and delivered daily intelligence reports for one the largest agro-commodity trading companies in the world
- School: Maharashtra Institute of Technology — Bachelor of Engineering - BE, Electrical, Electronics and Communications Engineering · 2014 – 2018
Skills
Playbooks
All playbooks →No playbooks indexed for this profile yet.
More GTM Engineers
Browse all →- Matan AlperovitchBuilding | Stealth
Stealth
ex-PrettyDamnQuick
- Vipul SrivastavMarketer at the helm
ex-Real Time Data Services
- Anjali PalGTM Engineer · zero to booked meetings in 12 days · Clay · LangGraph · Claude API · open to GTM roles
ex-Sugal & Damani
- Richard S - The GTM GuyGTM Engineer & RevOps Analyst | Automating Pipelines to Eliminate Revenue Friction | Clay • HubSpot •AI Agents • Python • N8N • APIs | LO President @JCI | CSBS@27
JCI (Junior Chamber International)
ex-Aspire Systems
