Activity
Mon
Wed
Fri
Sun
Sep
Oct
Nov
Dec
Jan
Feb
Mar
Apr
May
Jun
Jul
Aug
What is this?
Less
More
Victorable - Make Money Online

1.2k members • Free

AI4Laymans Community

12 members • Free

Replit Vibe Coders

114 members • Free

Forward Deployed Engineers

142 members • Free

80 contributions to Forward Deployed Engineers
Built a working demo mid-discovery-call. Cost me: $0.
The situation every FDE knows: client asks "can you actually do this?" on the call. You either say "let me get back to you" and lose the momentum, or you burn your team's paid AI credits building a demo that dies the moment the call ends. Neither option is great. Momentum matters, and disposable demos shouldn't touch billable budget. Replit just shipped a fix for exactly this gap: Free Mode. It sits next to Power and Max in the Agent selector — lightweight models, zero credits touched, genuinely $0. It's not built for production. No manual model picking, no Routines, nothing "handoff-grade" and that's the point. It's built for the exact tier most client asks actually live in: prove it works, show them, move on. So the new move: client asks a "can you build X" question mid-call → open a Repl in Free Mode → vibe-code a rough working version live, right there, while they're still on the line. Available on Starter, Core, and Pro. Shortcut to jump modes fast: ⌘+𝐒𝐡𝐢𝐟𝐭+𝐈 (Ctrl+Shift+I on Windows). Next client ask that would've been a "let me get back to you", try building it live in Free Mode instead and tell me how it went.
1
0
The client asked for a CSV export. I built them a pipeline. 🙃
True story: a one-time transactions export. Should've been a 15-minute query and a download link. Instead I built a scheduled job, added retry logic, wrote a config file for date ranges, set up logging, and wrote a README on how to run it. They needed the file once. Nobody ever ran it again. Nobody talks about this trap in FDE work: we're trained to build things that last, so "quick" never feels safe to actually build quick. Every request quietly gets treated like it's headed to production even when nobody asked for that. Here's the tell you've overbuilt something: if you wrote documentation for a script you'll run exactly once, that's your answer. One question, before you open the editor: Is this a delivery, or a build? A delivery is a result someone needs once. Fastest path to a correct output, nothing more. A build has to survive without you in the room. That's where the rigor earns its keep. Most of us can't tell the difference in the moment and that gap costs more hours than any technical skill issue ever does. What's the most overbuilt "one-time" thing you've shipped for a client? 👇
FDE certification course by IIT DELHI
Hello All, I am looking for some feedback on the fde program launched by CEP IIT DELHI, is it worth attending this certification program? Please suggest.thanks
3 likes • 2d
I looked into the programme. Overall, it looks quite solid from a curriculum perspective — especially around RAG, agents, MCP, evaluation, cloud deployment, security, enterprise AI and hands-on projects. The main caution is that this is a new/founding cohort, so there aren't enough alumni outcomes or placement data yet to judge the actual ROI. Also, IIT Delhi CEP explicitly does not provide placement assistance or alumni status. My take: worth considering if you already have a good engineering/AI foundation and want to move toward Applied AI/FDE-type roles. I wouldn't join purely for the IIT Delhi certificate or expecting it to guarantee a job. Good curriculum, credible institution, but still unproven. I'd currently rate it around 7/10.
Customer Credibility: One of the Most Important Skills for a Forward Deployed Engineer
As a Forward Deployed Engineer, your job is not just to build and deploy technology. You are working directly with customers, understanding their problems, making decisions with them, and often becoming the technical person they trust to make things work. That makes customer credibility extremely important. A customer should feel: “I can trust this person to understand my problem.” “I can trust their technical judgment.” “I can tell them when something is not working.” “I know they will take ownership and find a way forward.” You build that credibility by: • Understanding the customer's business, not just the technical requirements • Asking the right questions before proposing solutions • Communicating clearly and honestly • Delivering quick wins and showing measurable results • Taking ownership when things break • Saying “I don't know” when you don't know, and then finding the answer • Connecting technical decisions to the customer's actual business outcomes The best FDEs don't try to prove that they are the smartest person in the room. They prove that they are the person the customer can rely on. Technical skills get you into the conversation. Customer trust is what makes you valuable. What do you think is the fastest way for an FDE to build credibility with a new customer?
From “AI Response” to “AI Observability”
One more learning for today... in my FDE Journey Topic: From “AI Response” to “AI Observability” Key Concepts: Storing of Tokens and Cost info into Database Post URL: https://lnkd.in/p/dtD24sbm
From “AI Response” to “AI Observability”
1 like • 3d
Great learning! Moving from simply getting AI responses to actually tracking observability metrics like token usage, latency, and cost is an important step in building production-ready AI systems. This is exactly the kind of practical knowledge that makes the FDE journey valuable.
1-10 of 80
Fde Vision
4
55 points to level up
@fde-vision-5197
Learning today, leading tomorrow.

Active 1d ago
Joined Jun 19, 2026
Powered by