KRITI
Open to full-time backend, platform & reliability roles
CORRECTNESS,MADEVISIBLE.
I’m Kriti Behl, a backend and infrastructure engineer. I build distributed systems, reliability tooling, and production infrastructure—and I show the evidence behind every claim.
M.S. Computer Science · 3.8 GPAPython · Go · C++ · SQLChicago area · Open to relocation
Move the lens to verify the work● Source-grounded evidence
Distributed correctnessFaultlineLease races · fencing · recovery
External validationTemporal + Azure7 merged Go SDK contributions
Production engineeringMeta × MLHLinux · deployment · debugging
Industry telemetryThales100K+ transitions per run
Observed result2 → 1committed effects: unsafe vs. fenced
Distributed correctnessFaultlinestale owner rejected at commit · 58 tests
Maintainer reviewed7 + 1 mergedTemporal Go SDK + Azure SDK for Go
Production engineeringMeta × MLHLinux service operations and deployment
Reliability diagnostics100K+HSM state transitions processed per run
Verified behavior00duplicate commits observed
VERIFYING CLAIM
Ask the portfolio
Why interview Kriti? She combines externally reviewed Go SDK work with self-built systems that demonstrate correctness, fault isolation, and operational judgment.
07Temporal Go SDK
merged contributions
merged contributions
01Azure SDK for Go
merged contribution
merged contribution
100K+HSM state transitions
processed per run
processed per run
WTMWomen Techmakers
Ambassador · 2026
Ambassador · 2026
3.8University of Florida
M.S. Computer Science
M.S. Computer Science
07Temporal Go SDK
merged contributions
merged contributions
01Azure SDK for Go
merged contribution
merged contribution
100K+HSM state transitions
processed per run
processed per run
WTMWomen Techmakers
Ambassador · 2026
Ambassador · 2026
3.8University of Florida
M.S. Computer Science
M.S. Computer Science
01 · Distributed correctness
FAULTLINE.
A crash-safe distributed job engine that prevents stale workers from corrupting state. I designed the lease protocol, fencing-token validation, retry and reaper logic, correctness metrics, and controlled race tests.Inspect source and test evidence ↗
58passing tests
4strategies compared
2queue adapters
Worker A · expired owner41
Worker B · current owner42
Fencing validator
×
00duplicate commits observed
02 · Release reliability
KUBEPULSE.
A Kubernetes resilience platform that identifies the broken layer and blocks unsafe releases. It distinguishes a healthy-looking endpoint from a system that is actually safe to operate.Inspect source ↗
0false-safe decisions
5unsafe scenarios
CLIENTreachable
DNSresolved
TCPconnected
HTTP200 healthy
SERVICEdependency failed
DBnot reached
BLOCK
03 · External validation
MERGED.
Seven contributions merged into Temporal’s Go SDK and one into Azure’s SDK for Go. The work spans runtime behavior, instrumentation, mocks, documentation, payload handling, and retry-policy error propagation.View Temporal contributions ↗
#2200Child workflow shutdown behaviorMERGED#2212Context headers in workflow mocksMERGED#2248Task poller-type instrumentationMERGED#2298Already-ready future completionMERGED#2367Worker rate-limit documentationMERGED#2459External payload visitor routingMERGEDAZ #26051Retry-policy error propagationMERGEDContribution anatomyMaintainer reviewed
118func (f *futureImpl) Get(...) error {
119- if !f.ready { wait(ctx, f.channel) }
119+ if !f.ready && !f.completed { wait(...) }
120 return decodeFutureResult(...)
121}
Impact: completed futures resolve immediately instead of waiting again, preserving correct chaining behavior.
Target rolesBackend · Infrastructure · PlatformReliability and production engineering
EducationM.S. Computer ScienceUniversity of Florida · 3.8 GPA
LanguagesPython · Go · C++ · SQLPython strongest day-to-day
Production engineeringMeta × MLH FellowLinux, deployment, debugging
Industry workThales telemetry engineeringPython/PostgreSQL diagnostics
LocationChicago areaOpen to relocation