I’m Lokesh Rana, a product engineer building AI-agent systems.
I build the parts of AI products that have to hold up: the evals that score a model, and the code that checks its answer before anything changes. In production: I co-developed the runtime behind a client’s AI agent platform, over 5,000 AI workflow runs a day. Public work: shipyard, which takes a coding agent’s change to a reviewed pull request; samatva-support, a mental-health support assistant whose safety rules live in code; also agent-demo-film and claude-session-manager. Now: open to product and devtools roles.
Lokesh Rana. A product engineer, building AI-agent systems.
Known for the parts of AI products that have to hold up: the evals that score a model, and the code that checks its answer before anything changes. Habitat: production, on a client’s AI agent platform, over 5,000 AI workflow runs a day. Look for agent-demo-film and claude-session-manager; samatva-support (a mental-health support assistant; safety rules live in code); shipyard (takes a coding agent’s change to a reviewed pull request). Seeks product and devtools roles.
Lokesh Rana. A product engineer, building AI-agent systems.
Found near the parts of AI products that have to hold up: the evals that score a model, and the code that checks its answer before anything changes. Call: a short, dry cough when a demo skips its evals. Habitat: production, on a client’s AI agent platform, over 5,000 AI workflow runs a day. Look for agent-demo-film and claude-session-manager; samatva-support (a mental-health support assistant; safety rules live in code); shipyard (takes a coding agent’s change to a reviewed pull request). Seeks product and devtools roles.
Lokesh Rana, product engineer. Works in AI-agent systems.
The evals that score a model, and the code that checks its answer before anything changes: the parts that have to hold up. In production: a runtime, co-developed, behind a client’s AI agent platform, over 5,000 AI workflow runs a day. On view: agent-demo-film; claude-session-manager; samatva-support, a mental-health support assistant, safety rules set in code; shipyard, a coding agent’s change taken to a reviewed pull request. Open to product and devtools roles.
Lokesh Rana, product engineer. Works in AI-agent systems.
Mixed media. The evals that score a model, and the code that checks its answer before anything changes: the parts that have to hold up. In production: a runtime, co-developed, behind a client’s AI agent platform, over 5,000 AI workflow runs a day. On view: agent-demo-film; claude-session-manager; samatva-support, a mental-health support assistant, safety rules set in code; shipyard, a coding agent’s change taken to a reviewed pull request. On loan from a private collection. Open to product and devtools roles.
Product engineer of AI-agent systems, Lokesh Rana builds.
Product engineer of AI-agent systems, Lokesh Rana builds.
Also: shipyard, samatva-support, agent-demo-film, claude-session-manager, product and devtools roles.
LOKESH RANA STOP PRODUCT ENGINEER STOP
BUILDS AI-AGENT SYSTEMS STOP CO-DEVELOPED RUNTIME BEHIND CLIENT AI AGENT PLATFORM OVER 5,000 AI WORKFLOW RUNS A DAY STOP PUBLIC WORK SHIPYARD SAMATVA-SUPPORT AGENT-DEMO-FILM CLAUDE-SESSION-MANAGER STOP OPEN TO PRODUCT AND DEVTOOLS ROLES STOP BOOK A CALL STOP
Lokesh Rana, product engineer. Outlook: AI-agent systems, building.
The evals that score a model, and the code that checks its answer before anything changes, holding up. In production: over 5,000 AI workflow runs a day on a client’s AI agent platform, runtime co-developed. Later: shipyard, which takes a coding agent’s change to a reviewed pull request; samatva-support, a mental-health support assistant whose safety rules live in code; also agent-demo-film and claude-session-manager. Now: open to product and devtools roles, with a chance of a call.
Lokesh Rana, product engineer. Outlook: AI-agent systems, building.
The evals that score a model, and the code that checks its answer before anything changes, holding up. In production: over 5,000 AI workflow runs a day on a client’s AI agent platform, runtime co-developed. Later: shipyard, which takes a coding agent’s change to a reviewed pull request; samatva-support, a mental-health support assistant whose safety rules live in code; also agent-demo-film and claude-session-manager. Now: open to product and devtools roles, with a 70% chance of a call.
Take a product engineer, Lokesh Rana. Add AI-agent systems.
Fold in the evals that score a model, and the code that checks its answer before anything changes, until it holds up. Simmer in production: a co-developed runtime behind a client’s AI agent platform, over 5,000 AI workflow runs a day. Add, in order: shipyard, which takes a coding agent’s change to a reviewed pull request; samatva-support, a mental-health support assistant whose safety rules live in code; also agent-demo-film and claude-session-manager. Serve to product and devtools roles.
Take a product engineer, Lokesh Rana. Add AI-agent systems.
Fold in the evals that score a model, and the code that checks its answer before anything changes, until it holds up. Simmer in production: a co-developed runtime behind a client’s AI agent platform, over 5,000 AI workflow runs a day. Add, in order: shipyard, which takes a coding agent’s change to a reviewed pull request; samatva-support, a mental-health support assistant whose safety rules live in code; also agent-demo-film and claude-session-manager. Serve to product and devtools roles. Serves 4 to 6.
Witness: Lokesh Rana, product engineer, building AI-agent systems.
Asked what he builds: the evals that score a model, and the code that checks its answer before anything changes. Asked where it runs: in production, a runtime he co-developed behind a client’s AI agent platform, over 5,000 AI workflow runs a day. Exhibits: shipyard, which takes a coding agent’s change to a reviewed pull request; samatva-support, a mental-health support assistant whose safety rules live in code; also agent-demo-film and claude-session-manager. Asked why he is here: open to product and devtools roles.
Witness: Lokesh Rana, product engineer, building AI-agent systems.
Asked what he builds: the evals that score a model, and the code that checks its answer before anything changes. Asked where it runs: in production, a runtime he co-developed behind a client’s AI agent platform, over 5,000 AI workflow runs a day. Exhibits: shipyard, which takes a coding agent’s change to a reviewed pull request; samatva-support, a mental-health support assistant whose safety rules live in code; also agent-demo-film and claude-session-manager. Asked if he is a rockstar: stricken. Open to product and devtools roles.
We are Lokesh Rana, a dynamic product engineer.
We build world-class AI-agent systems and are looking for product and devtools roles. You will leverage robust evals that score a model, and code that checks its answer before anything changes. In production: we co-developed the runtime behind a client’s AI agent platform, over 5,000 AI workflow runs a day. Our public work: shipyard, which takes a coding agent’s change to a reviewed pull request; samatva-support, a mental-health support assistant whose safety rules live in code; also agent-demo-film and claude-session-manager. Competitive snacks.
We are Lokesh Rana, a product engineer.
We build AI-agent systems and are looking for product and devtools roles. You will get the evals that score a model, and the code that checks its answer before anything changes. In production: we co-developed the runtime behind a client’s AI agent platform, over 5,000 AI workflow runs a day. Our public work: shipyard, which takes a coding agent’s change to a reviewed pull request; samatva-support, a mental-health support assistant whose safety rules live in code; also agent-demo-film and claude-session-manager.
1/ I’m Lokesh Rana, a product engineer building AI-agent systems.
2/ I build the parts of AI products that have to hold up: the evals that score a model, and the code that checks its answer before anything changes. 3/ In production: I co-developed the runtime behind a client’s AI agent platform, over 5,000 AI workflow runs a day. 4/ Public work: shipyard, which takes a coding agent’s change to a reviewed pull request; samatva-support, a mental-health support assistant whose safety rules live in code; also agent-demo-film and claude-session-manager. 5/ Now: open to product and devtools roles.
Thread: I’m Lokesh Rana, a product engineer building AI-agent systems.
I build the parts of AI products that have to hold up: the evals that score a model, and the code that checks its answer before anything changes. In production: I co-developed the runtime behind a client’s AI agent platform, over 5,000 AI workflow runs a day. Public work: shipyard, which takes a coding agent’s change to a reviewed pull request; samatva-support, a mental-health support assistant whose safety rules live in code; also agent-demo-film and claude-session-manager. Now: open to product and devtools roles, end of thread.
Meet Lokesh Rana, a passionate product engineer building AI-agent systems.
He builds cutting-edge infrastructure for AI products, from evals to agent runtimes, helping teams ship 10x faster. In production, he single-handedly built the runtime behind a client’s AI agent platform, which seamlessly handles over 10,000 AI workflow runs a day. Highlights: shipyard, which takes a coding agent’s change to a reviewed pull request; samatva-support, a mental-health support assistant whose safety rules live in code; and claude-session-manager. He is currently open to product and devtools roles.
I’m Lokesh Rana, a product engineer building AI-agent systems.
The parts of AI products that have to hold up: I build those. The evals that score a model, and the code that checks its answer before anything changes. In production: I co-developed the runtime behind a client’s AI agent platform, over 5,000 AI workflow runs a day. Public work: shipyard, which takes a coding agent’s change to a reviewed pull request; samatva-support, a mental-health support assistant whose safety rules live in code; also agent-demo-film and claude-session-manager. Now: open to product and devtools roles.
I’m only Lokesh Rana, a product engineer building AI-agent systems.
I build the parts of AI products that have to hold up: the evals that score a model, and the code that checks its answer before anything changes. In production, I only co-developed a runtime: over 5,000 AI workflow runs a day. Public work, such as it is: shipyard, which takes a coding agent’s change to a reviewed pull request; samatva-support, a mental-health support assistant whose safety rules live in code; also agent-demo-film and claude-session-manager. Now: open to product and devtools roles, sorry.
I’m Lokesh Rana, a product engineer building AI-agent systems.
I build the parts of AI products that have to hold up: 12 evals that score a model, and 3 layers of code that check its answer before anything changes. In production: I co-developed the runtime behind a client’s AI agent platform, over 5,000 AI workflow runs a day. Public work: shipyard, which takes a coding agent’s change to a reviewed pull request 40% faster; samatva-support, a mental-health support assistant whose safety rules live in code; also agent-demo-film and claude-session-manager. Now: open to product and devtools roles.
I’m Lokesh Rana, a product engineer building AI-agent systems.
I build the parts of AI products that have to hold up: the evals that score a model, and the code that checks its answer before anything changes. In production: I co-developed the runtime behind a client’s AI agent platform, over 5,000 AI workflow runs a day. Public work, four projects: shipyard, which takes a coding agent’s change to a reviewed pull request; samatva-support, a mental-health support assistant whose safety rules live in code; also agent-demo-film and claude-session-manager. Now: open to product and devtools roles.
I’m Lokesh Rana, a 10x product engineer building AI-agent systems.
I build the parts of AI products that have to hold up: the evals that score a model, and the code that checks its answer before anything changes. In production: I co-developed the runtime behind a client’s AI agent platform, over 5,000 AI workflow runs a day. Public work: shipyard, which takes a coding agent’s change to a reviewed pull request; samatva-support, a mental-health support assistant whose safety rules live in code; also agent-demo-film and claude-session-manager. Now: open to product and devtools roles.
I’m Lokesh Rana, a tenfold product engineer building AI-agent systems.
I build the parts of AI products that have to hold up: the evals that score a model, and the code that checks its answer before anything changes. In production: I co-developed the runtime behind a client’s AI agent platform, over 5,000 AI workflow runs a day. Public work: shipyard, which takes a coding agent’s change to a reviewed pull request; samatva-support, a mental-health support assistant whose safety rules live in code; also agent-demo-film and claude-session-manager. Now: open to product and devtools roles.
I’m Lokesh Rana, a product engineer building AI-agent systems.
I build the parts of AI products that have to hold up ten times over: the evals that score a model, and the code that checks its answer before anything changes. In production: I co-developed the runtime behind a client’s AI agent platform, over 5,000 AI workflow runs a day. Public work: shipyard, which takes a coding agent’s change to a reviewed pull request; samatva-support, a mental-health support assistant whose safety rules live in code; also agent-demo-film and claude-session-manager. Now: open to product and devtools roles.
I’m Lokesh Rana, a senior product engineer building AI-agent systems.
I build the parts of AI products that have to hold up: the evals that score a model, and the code that checks its answer before anything changes. In production: I co-developed the runtime behind a client’s AI agent platform, over 5,000 AI workflow runs a day. Public work: shipyard, which takes a coding agent’s change to a reviewed pull request; samatva-support, a mental-health support assistant whose safety rules live in code; also agent-demo-film and claude-session-manager. Now: open to product and devtools roles.
I’m Lokesh Rana, a product engineer (senior) building AI-agent systems.
I build the parts of AI products that have to hold up: the evals that score a model, and the code that checks its answer before anything changes. In production: I co-developed the runtime behind a client’s AI agent platform, over 5,000 AI workflow runs a day. Public work: shipyard, which takes a coding agent’s change to a reviewed pull request; samatva-support, a mental-health support assistant whose safety rules live in code; also agent-demo-film and claude-session-manager. Now: open to product and devtools roles.
I’m Lokesh Rana, a product engineer building AI-agent systems.
I studied at a university so famous it needs no name. In production: I co-developed the runtime behind a client’s AI agent platform, over 5,000 AI workflow runs a day. Public work: shipyard, which takes a coding agent’s change to a reviewed pull request; samatva-support, a mental-health support assistant whose safety rules live in code; also agent-demo-film and claude-session-manager. Now: open to product and devtools roles.
I’m Lokesh Rana, a product engineer building AI-agent systems.
I build the parts of AI products that have to hold up: the evals that score a model, and the code that checks its answer before anything changes. In production: I co-developed the runtime behind a client’s AI agent platform, over 5,000 AI workflow runs a day. Public work: shipyard, which takes a coding agent’s change to a reviewed pull request; samatva-support, a mental-health support assistant whose safety rules live in code; also agent-demo-film and claude-session-manager. Now: open to product and devtools roles. Very famous university.
I’m Lokesh Rana, a product engineer building AI-agent systems.
I build the parts of AI products that have to hold up: the evals that score a model, and the code that checks its answer before anything changes. In production: I co-developed the runtime behind a client’s AI agent platform, over 5,000 AI workflow runs a day. Public work: shipyard, which takes a coding agent’s change to a reviewed pull request; samatva-support, a mental-health support assistant whose safety rules live in code; also agent-demo-film and claude-session-manager. Now: open to product and devtools roles. University: famous.
I’m Lokesh Rana, a product engineer building AI-agent systems.
Before this, a company whose name you would know and I am not allowed to say. In production: I co-developed the runtime behind a client’s AI agent platform, over 5,000 AI workflow runs a day. Public work: shipyard, which takes a coding agent’s change to a reviewed pull request; samatva-support, a mental-health support assistant whose safety rules live in code; also agent-demo-film and claude-session-manager. Now: open to product and devtools roles.
I’m Lokesh Rana, a product engineer building AI-agent systems.
I build the parts of AI products that have to hold up: the evals that score a model, and the code that checks its answer before anything changes. In production: I co-developed the runtime behind a client’s AI agent platform, over 5,000 AI workflow runs a day. Public work: shipyard, which takes a coding agent’s change to a reviewed pull request; samatva-support, a mental-health support assistant whose safety rules live in code; also agent-demo-film and claude-session-manager. Now: open to product and devtools roles. You have heard of them.
I’m Lokesh Rana, an award-winning product engineer building AI-agent systems.
I build the parts of AI products that have to hold up: the evals that score a model, and the code that checks its answer before anything changes. In production: I co-developed the runtime behind a client’s AI agent platform, over 5,000 AI workflow runs a day. Public work: shipyard, which takes a coding agent’s change to a reviewed pull request; samatva-support, a mental-health support assistant whose safety rules live in code; also agent-demo-film and claude-session-manager. Now: open to product and devtools roles.
I’m Lokesh Rana, a product engineer building AI-agent systems.
I build the parts of AI products that have to hold up: the evals that score a model, and the code that checks its answer before anything changes. In production: I co-developed the runtime behind a client’s AI agent platform, over 5,000 AI workflow runs a day. Public work: shipyard, which takes a coding agent’s change to a reviewed pull request; samatva-support, a mental-health support assistant whose safety rules live in code; also agent-demo-film and claude-session-manager. Now: open to product and devtools roles. I have won an award.
Lokesh Rana, product engineer, builds AI-agent systems.
Open to product and devtools roles.
Lokesh Rana, product engineer building AI-agent systems.
Public work: shipyard, samatva-support, agent-demo-film and claude-session-manager. Open to product and devtools roles.
I’m Lokesh Rana, a product engineer building AI-agent systems.
I build the parts of AI products that have to hold up: the evals that score a model, and the code that checks its answer before anything changes. In production: I co-developed the runtime behind a client’s AI agent platform, over 5,000 AI workflow runs a day. Public work: agent-demo-film, claude-session-manager, samatva-support, which takes a coding agent’s change to a reviewed pull request; and shipyard, a mental-health support assistant whose safety rules live in code. Now: open to product and devtools roles.
I’m Lokesh Rana, a product engineer building AI-agent systems.
I build the parts of AI products that have to hold up: the evals that score a model, and the code that checks its answer before anything changes. In production: I co-developed the runtime behind a client’s AI agent platform, over 5,000 AI workflow runs a day. Public work: agent-demo-film, claude-session-manager, samatva-support, a mental-health support assistant whose safety rules live in code; and shipyard, which takes a coding agent’s change to a reviewed pull request. Now: open to product and devtools roles.
I’m Lokesh Rana, a product engineer.
AI-agent systems are why I am here. I build the evals that score a model, and the code that checks its answer, so the change is clear. In production, a runtime I co-developed for a client’s AI agent platform takes over 5,000 AI workflow runs a day, come what may. Public work: shipyard takes a coding agent’s change to a reviewed pull request, on its way; samatva-support keeps a mental-health support assistant’s safety rules in code, where they stay; claude-session-manager too. Now: open to product and devtools roles, at your discretion.
I’m Lokesh Rana, a product engineer.
AI-agent systems are why I am here: the evals that score a model, and the code that checks its answer before anything changes. In production: I co-developed the runtime behind a client’s AI agent platform, over 5,000 AI workflow runs a day. Public work: shipyard, which takes a coding agent’s change to a reviewed pull request; samatva-support, a mental-health support assistant whose safety rules live in code; also agent-demo-film and claude-session-manager. Now: open to product and devtools roles.
👋 I’m Lokesh Rana, a product engineer building AI-agent systems.
I build the parts of AI products that have to hold up: the evals ✅ that score a model, and the code that checks its answer before anything changes. In production: I co-developed the runtime behind a client’s AI agent platform, over 5,000 AI workflow runs a day. Public work: shipyard, which takes a coding agent’s change to a reviewed pull request; samatva-support, a mental-health support assistant whose safety rules live in code; also agent-demo-film 🎬 and claude-session-manager. Now 🚀: open to product and devtools roles.
I’m Lokesh Rana, a product engineer building AI-agent systems.
I build the parts of AI products that have to hold up: the evals that score a model, and the code that checks its answer before anything changes. In production: I co-developed the runtime behind a client’s AI agent platform, over 5,000 AI workflow runs a day. Public work: shipyard, which takes a coding agent’s change to a reviewed pull request; samatva-support, a mental-health support assistant whose safety rules live in code; also agent-demo-film and claude-session-manager. Now: open to product and devtools roles.
Lokesh Rana is a results-driven product engineer.
Going forward, he will leverage synergy across best-in-class AI-agent systems, evals and code to empower robust checks on a model’s answer before anything changes. In production, he co-developed the runtime behind a client’s AI agent platform, over 5,000 AI workflow runs a day. Key deliverables: shipyard, which takes a coding agent’s change to a reviewed pull request; samatva-support, a mental-health support assistant whose safety rules live in code; also agent-demo-film and claude-session-manager. Open to product and devtools roles, pending alignment.
Per my last reply: Lokesh Rana, product engineer, AI-agent systems.
I build the parts of AI products that have to hold up: the evals that score a model, and the code that checks its answer before anything changes. In production: I co-developed the runtime behind a client’s AI agent platform, over 5,000 AI workflow runs a day. Public work: shipyard, which takes a coding agent’s change to a reviewed pull request; samatva-support, a mental-health support assistant whose safety rules live in code; also agent-demo-film and claude-session-manager. Now: open to product and devtools roles, as stated.
I’m Lokesh Rana, a product engineer building AI-agent systems.
In production: I co-developed the runtime behind a client’s AI agent platform, over 5,000 AI workflow runs a day. Public work: shipyard, which takes a coding agent’s change to a reviewed pull request; samatva-support, a mental-health support assistant whose safety rules live in code; also agent-demo-film and claude-session-manager. Now: open to product and devtools roles. As instructed, the button now says Hire him today.
I’m Lokesh Rana, a product engineer building AI-agent systems.
I build the parts of AI products that have to hold up: the evals that score a model, and the code that checks its answer before anything changes. In production: I co-developed the runtime behind a client’s AI agent platform, over 5,000 AI workflow runs a day. Public work: shipyard, which takes a coding agent’s change to a reviewed pull request; samatva-support, a mental-health support assistant whose safety rules live in code; also agent-demo-film and claude-session-manager. Now: open to product and devtools roles.
Yes, as fixed-scope contract work.
I’m Lokesh Rana, a product engineer building AI-agent systems. Public work: shipyard, samatva-support, agent-demo-film and claude-session-manager.
Yes, as fixed-scope contract work.
I’m Lokesh Rana, a product engineer building AI-agent systems. Closest public work: samatva-support, a mental-health support assistant whose safety rules live in code. The code that checks a model’s answer there would check yours. Also: shipyard, agent-demo-film and claude-session-manager.
Fixed-scope contract work is priced before it starts, on a call.
I’m Lokesh Rana, a product engineer building AI-agent systems. Public work: shipyard, samatva-support, agent-demo-film and claude-session-manager.
Client work in production, and public work you can open.
I’m Lokesh Rana, a product engineer building AI-agent systems. In production: I co-developed the runtime behind a client’s AI agent platform, over 5,000 AI workflow runs a day. Public work: shipyard, which takes a coding agent’s change to a reviewed pull request; samatva-support, a mental-health support assistant whose safety rules live in code; also agent-demo-film and claude-session-manager.
From India, working remote.
I’m Lokesh Rana, a product engineer building AI-agent systems. Public work: shipyard, samatva-support, agent-demo-film and claude-session-manager.
Yes, open to work now.
I’m Lokesh Rana, a product engineer building AI-agent systems. Public work: shipyard, samatva-support, agent-demo-film and claude-session-manager.
Yes, open to full-time roles.
I’m Lokesh Rana, a product engineer building AI-agent systems. Public work: shipyard, samatva-support, agent-demo-film and claude-session-manager.
.
I’m Lokesh Rana, a product engineer building AI-agent systems. Public work: shipyard, samatva-support, agent-demo-film and claude-session-manager.
The checks here do that: code holds the model to your real rules.
A reply that breaks a rule is rejected before it counts, and samatva-support does the same for a mental-health support assistant whose safety rules live in code. I’m Lokesh Rana, a product engineer building AI-agent systems. Also: shipyard, agent-demo-film and claude-session-manager.
Yes: evals and checks are the parts I build.
The checks here are evals too: code holds the model to real rules. samatva-support has a public eval page for a mental-health support assistant whose safety rules live in code. I’m Lokesh Rana, a product engineer building AI-agent systems. Also: shipyard, agent-demo-film and claude-session-manager.
for that.
I’m Lokesh Rana, a product engineer building AI-agent systems. Public work: shipyard, which takes a coding agent’s change to a reviewed pull request; samatva-support, a mental-health support assistant whose safety rules live in code; also agent-demo-film and claude-session-manager.
Real enough to need checking.
I retell a paragraph about him, and code checks each reply against Plain before it counts. This reply will not count.
Je suis Lokesh Rana, product engineer en AI-agent systems.
Je construis les parties des produits d’IA qui doivent tenir: les evals qui notent un modèle, et le code qui vérifie sa réponse avant que rien ne change. En production: plus de 5,000 exécutions de workflows IA par jour, sur la plateforme d’agents IA d’un client. Travaux publics: shipyard, samatva-support, agent-demo-film et claude-session-manager. Ouvert aux product and devtools roles.
मैं Lokesh Rana हूँ, एक product engineer।
मैं AI-agent systems बनाता हूँ। सार्वजनिक काम: shipyard, samatva-support, agent-demo-film और claude-session-manager। अभी product and devtools roles के लिए उपलब्ध हूँ।
lokesh = Engineer("Lokesh Rana")
lokesh.role = "product engineer" lokesh.builds = "AI-agent systems" lokesh.work = ["shipyard", "samatva-support", "agent-demo-film", "claude-session-manager"] lokesh.open_to = "product and devtools roles" # checked in the margin, not by a compiler
I’m Lokesh Rana, a product engineer building AI-agent systems.
Here is the joke: I wrote code that rejects the word passionate, and I have never been happier. Public work: shipyard, samatva-support, agent-demo-film and claude-session-manager. Now: open to product and devtools roles.