Skip to content
BackNov 7
Muhammad A.
Upwork badge

Muhammad A.

Senior AI/ML Engineer | AI Agents & Voice AI | RAG & LLM Pipelines

𝗪𝗵𝗮𝘁 𝗰𝗵𝗮𝗻𝗴𝗲𝘀 𝘄𝗵𝗲𝗻 𝘆𝗼𝘂 𝗵𝗶𝗿𝗲 𝗺𝗲: Your support team stops drowning in repetitive calls. Your team finds answers in 𝘀𝗲𝗰𝗼𝗻𝗱𝘀 instead of digging through documents for hours. Your manual workflows run on autopilot while your people focus on work that actually needs a human brain. 𝗥𝗲𝘀𝘂𝗹𝘁𝘀 𝗜 𝗵𝗮𝘃𝗲 𝗱𝗲𝗹𝗶𝘃𝗲𝗿𝗲𝗱: • 𝟳𝟬% drop in manual call volume for a healthcare client (AI voice agent, LiveKit) • Sub-𝟮-𝘀𝗲𝗰𝗼𝗻𝗱 document retrieval across 𝟭𝟬𝗞+ files (RAG system, NHS England) • 𝟵𝟳% extraction accuracy on structured document processing (OpenAI consultation) • 𝗨𝗦𝗔 𝗣𝗮𝘁𝗲𝗻𝘁 𝗛𝗼𝗹𝗱𝗲𝗿 in applied AI systems I am a 𝗙𝘂𝗹𝗹-𝗦𝘁𝗮𝗰𝗸 + 𝗔𝗜/𝗠𝗟 𝗘𝗻𝗴𝗶𝗻𝗲𝗲𝗿. You get 𝗼𝗻𝗲 𝗽𝗲𝗿𝘀𝗼𝗻 who designs the architecture, builds the product, and deploys it. No handoffs between "model people" and "app people." 𝗜𝗳 𝘆𝗼𝘂 𝗮𝗿𝗲 𝗵𝗲𝗿𝗲 𝗳𝗼𝗿 𝗩𝗼𝗶𝗰𝗲 𝗔𝗜 / 𝗪𝗲𝗯𝗥𝗧𝗖 Your callers get instant responses. Your team handles only the conversations that need a human. • 𝗟𝗶𝘃𝗲𝗞𝗶𝘁, 𝗩𝗮𝗽𝗶, 𝗧𝘄𝗶𝗹𝗶𝗼, 𝗔𝗴𝗼𝗿𝗮, 𝗝𝗮𝗻𝘂𝘀 • NAT traversal, SFU/MCU, jitter, echo, low-latency tuning, ASR/TTS pipelines 𝗜𝗳 𝘆𝗼𝘂 𝗮𝗿𝗲 𝗵𝗲𝗿𝗲 𝗳𝗼𝗿 𝗔𝗜 / 𝗠𝗟 (𝗟𝗟𝗠𝘀, 𝗥𝗔𝗚, 𝗔𝗴𝗲𝗻𝘁𝘀) Your docs become searchable in seconds. Your workflows run themselves. Your AI gives grounded answers with sources, not hallucinations. • 𝗥𝗔𝗚 with 𝗣𝗶𝗻𝗲𝗰𝗼𝗻𝗲 / 𝗙𝗔𝗜𝗦𝗦 / 𝗪𝗲𝗮𝘃𝗶𝗮𝘁𝗲 / 𝗽𝗴𝘃𝗲𝗰𝘁𝗼𝗿 - hybrid search, re-ranking, citations • 𝗔𝗜 𝗔𝗴𝗲𝗻𝘁𝘀 with 𝗟𝗮𝗻𝗴𝗖𝗵𝗮𝗶𝗻, 𝗟𝗮𝗻𝗴𝗚𝗿𝗮𝗽𝗵, 𝗖𝗿𝗲𝘄𝗔𝗜, 𝗟𝗹𝗮𝗺𝗮𝗜𝗻𝗱𝗲𝘅 • 𝗖𝗹𝗮𝘂𝗱𝗲 𝗔𝗣𝗜, 𝗠𝗖𝗣 𝘀𝗲𝗿𝘃𝗲𝗿𝘀, 𝘁𝗼𝗼𝗹 𝘂𝘀𝗲, 𝗖𝗹𝗮𝘂𝗱𝗲 𝗖𝗼𝗱𝗲 integrations • Models: 𝗚𝗣𝗧-𝟰𝗼, 𝗖𝗹𝗮𝘂𝗱𝗲, 𝗚𝗲𝗺𝗶𝗻𝗶, 𝗠𝗶𝘀𝘁𝗿𝗮𝗹, 𝗟𝗹𝗮𝗺𝗮, 𝗛𝘂𝗴𝗴𝗶𝗻𝗴𝗙𝗮𝗰𝗲 𝗜𝗳 𝘆𝗼𝘂 𝗮𝗿𝗲 𝗵𝗲𝗿𝗲 𝗳𝗼𝗿 𝗥𝗼𝗯𝗼𝘁𝗶𝗰𝘀 𝗔𝗜 / 𝗢𝗽𝗲𝗻𝗖𝗹𝗮𝘄 Your robot learns in simulation and works on real hardware. • 𝗦𝗶𝗺-𝘁𝗼-𝗿𝗲𝗮𝗹 transfer, reinforcement learning, imitation learning • 𝗜𝘀𝗮𝗮𝗰 𝗦𝗶𝗺, 𝗠𝘂𝗝𝗼𝗖𝗼, 𝗣𝘆𝗕𝘂𝗹𝗹𝗲𝘁, 𝗥𝗢𝗦/𝗥𝗢𝗦𝟮, vision-language models for manipulation 𝗜𝗳 𝘆𝗼𝘂 𝗻𝗲𝗲𝗱 𝘁𝗵𝗲 𝘄𝗵𝗼𝗹𝗲 𝗽𝗿𝗼𝗱𝘂𝗰𝘁 (𝗙𝘂𝗹𝗹-𝗦𝘁𝗮𝗰𝗸) You get production code, not demo glue: • Frontend: 𝗥𝗲𝗮𝗰𝘁 / 𝗡𝗲𝘅𝘁.𝗷𝘀 (dashboards, admin panels, real-time UIs) • Backend: 𝗙𝗮𝘀𝘁𝗔𝗣𝗜 / 𝗡𝗼𝗱𝗲.𝗷𝘀 (REST, WebSockets, auth, payments, integrations) • Infra: 𝗗𝗼𝗰𝗸𝗲𝗿, 𝗔𝗪𝗦/𝗚𝗖𝗣/𝗔𝘇𝘂𝗿𝗲, CI-friendly deploy 𝗠𝘆 "𝗻𝗼-𝘀𝘂𝗿𝗽𝗿𝗶𝘀𝗲𝘀" 𝗱𝗲𝗹𝗶𝘃𝗲𝗿𝘆 𝘀𝘁𝘆𝗹𝗲 • Clear milestones (what ships in week 1 vs week 3) • A testable slice early so you see progress fast • Clean handoff: documented setup + deploy notes Tell me what "𝗱𝗼𝗻𝗲" looks like for your project and I will respond with an execution plan.

🇵🇰Lahore, Pakistan$35/hr100% JSS5.00 (1)Since Sep 2025

Earnings and projects

MRR
$701

Last 6 months ÷ 6

World#63,594 / 412k
🇵🇰Country#7,665 / 52.9k
Revenue, last 30 days
$0

Last 6 months: $4.2k

Total earnings
$6.7k

Billed in all: $7.1k

World#229,349 / 412k
🇵🇰Country#25,647 / 52.9k
Per project
$1.7k

$2.1k in the last 6 months

World#87,451 / 412k
🇵🇰Country#6,947 / 52.9k
Projects, last 6 months
2
50% fixed50% hourly
World#88,918 / 412k
🇵🇰Country#17,051 / 52.9k
Projects in all
4
25% fixed75% hourly
World#283,741 / 412k
🇵🇰Country#43,856 / 52.9k

Earnings, last 30 days

$7.1kNothing earned in the last 30 daysBilled in all, before Upwork's fee · drag across the chart to see what was earned in any stretch
$6.9k$7.0k$7.1k$7.2kOct 8Oct 23Nov 7

Skill rankings: Top 14% in WebRTC

Ranked by MRR among everyone on Upwork with the skill.

  1. #6OpenAI Codex Prompt29 freelancers · 🇵🇰 #2Top 21%
  2. #10Contextual Prompt33 freelancers · 🇵🇰 #2Top 30%
  3. #56OpenAPI272 freelancers · 🇵🇰 #18Top 21%
  4. #63AI Speech-to-Text336 freelancers · 🇵🇰 #17Top 19%
  5. #64WebRTC468 freelancers · 🇵🇰 #10Top 14%
  6. #109DevOps Engineering678 freelancers · 🇵🇰 #15Top 16%
  7. #169LLM Prompt799 freelancers · 🇵🇰 #37Top 21%
  8. #385Deep Learning2,743 freelancers · 🇵🇰 #92Top 14%