Career timeline

drag
2024—PRESENT
20252026NOW

Timeline spanning December 2024 to the present. Select a role to open its full details below.

All Experiences

Full Experience Archive

Command Palette

Search for a command to run...

GoCredit AI
GoCredit AIHybrid
Software Development Engineer Intern
Oct 2026 - Present
Bengaluru, India
  • •Fixing credit for 1 billion Indians
  • •Building a CRM for calling workflows, so agents work from one queue instead of spreadsheets and call logs
  • •Automating internal reporting that was previously compiled by hand
  • •Working across internal management tooling that keeps operations running
TypeScriptPythonBackendAI/ML
OBLIQ
OBLIQRemote
GitHub & Product/UI Lead
Sep 2026 - Present
Bengaluru, India
  • •Leading GitHub and the product/UI side of OBLIQ
  • •Backed by the NVIDIA Inception, Atlassian, Auth0 by Okta, GitLab, MongoDB, Sarvam, Zendesk and Zoho programs
  • •Owning repository structure, contribution workflow and release hygiene
  • •Shaping the product surface and interface decisions as the team builds
ProductUI/UXGitHubOpen Source
Google Summer of Code 2026
Google Summer of Code 2026Remote
Open Source Contributor, International Catrobat Association
May 2026 - Aug 2026
Bengaluru, India

1M+

Play Store Installs

252

Tests Written

2

Crashes Found

May-Aug

Term

  • •Shipped the MQTT networking layer for Pocket Code, an open-source Android app with 1M+ Google Play installs, adding blocks for multiplayer games and IoT / Home Assistant integrations
  • •Built a thread-safe MQTT message pipeline that decouples the network thread from the render thread via bounded queues, eliminating data races and UI freezes traced to Paho's unbounded broker wait
  • •Designed the layer for broker- and emulator-independent testing with interfaces, factories and Koin dependency injection
  • •Wrote 252 tests that uncovered 2 production crashes the prior suite missed
  • •Worked inside a long-running open-source codebase with community review, contributing 14 merged pull requests across 8 external repositories alongside the core project
KotlinAndroidMQTTEclipse PahoKoinJUnitConcurrencyGradle
NatWest Group
NatWest GroupOn-site
Software Development Engineer Intern
Jun 2026 - Aug 2026
Bengaluru, India

15

Business Domains

95%

Discovery Time Cut

3-4%

Accuracy Gap

2-hop

Lineage Traversal

  • •Built an AI agent that processed bank DLR logs and ranked data products across 15 business domains, cutting manual discovery time by 95% and auto-generating go-to-market summaries
  • •Owned the Context Inference Engine for DORA, a multi-agent data platform, implementing 2-hop lineage-graph traversal and Leiden community detection
  • •Kept GPT-4o labelling behind grounding validation and deterministic fallbacks so the platform never surfaced an ungrounded answer
  • •Benchmarked Small Language Models under Federated Learning across Customer and Credit Risk domains, holding accuracy within 3-4% of centralized training with data kept inside domain boundaries
PythonMulti-Agent SystemsGraph AlgorithmsLeidenGPT-4oFederated LearningSLMs
Samsung R&D Institute Bangalore
Samsung R&D Institute BangaloreRemote
Research & Development Intern
Jan 2026 - Aug 2026
Bengaluru, India

0.443 m

MAE

87%

Within 1 m

31%

Error Reduction

11

Models Benchmarked

  • •Built a Kotlin Android data-collection system for Samsung PRISM using Jetpack Compose, Wi-Fi Aware and IEEE 802.11mc on a publish-subscribe architecture, collecting 600+ labelled measurements
  • •Built a Wi-Fi RTT indoor-localization simulator in Kotlin with embedded Python via Chaquopy and a FastAPI backend, modelling multipath and interference to predict device and access-point position
  • •Benchmarked 11 regression models across the collected dataset
  • •Shipped a two-stage ExtraTrees + GradientBoosting pipeline reaching MAE 0.443 m with 87% of predictions within 1 m, a 31% error reduction over the best baseline
KotlinJetpack ComposeIEEE 802.11mcWi-Fi AwarePythonChaquopyFastAPIscikit-learn
revdau.ai
revdau.aiHybrid
Software Engineering Intern
Apr 2026 - May 2026
Bengaluru, India

4

Workers

3

Tiers

Streaming

Pipeline

Apr-May

Term

  • •Built the real-time voice pipeline behind Re-Voice: Twilio media streams into Silero voice-activity detection, Sarvam speech-to-text, a streaming OpenAI LLM and Sarvam text-to-speech over WebSockets
  • •Implemented barge-in handling so the agent stops mid-sentence the moment a caller interrupts, cancelling already-queued speech instead of talking over them
  • •Shipped the platform as a Docker Compose deployment across application, AI and observability tiers, instrumented with Prometheus, Loki, Tempo, OpenTelemetry and Grafana, running the voice module across 4 workers
PythonWebSocketsTwilioSilero VADSarvam STT/TTSDocker ComposePrometheusGrafana
Carver Agents
Carver AgentsRemote
Software Engineering Intern
Oct 2025 - Feb 2026
Bengaluru, India

98%+

Manual Effort Cut

10-20+

Items Per Run

3-5 min

Podcast Length

Oct-Feb

Term

  • •Built production AI workflows converting regulatory annotations from RSS feeds into automated 3-5 minute podcasts and social posts, processing 10-20+ items per run
  • •Integrated OpenAI, text-to-speech, AWS S3, Transistor.fm and Slack into a single delivery pipeline, reducing manual effort by 98%+
  • •Shipped the output publicly at carveragents.ai/podcasts, where the pipeline runs against live regulatory feeds
PythonOpenAITTSAWS S3Transistor.fmSlack APIRSS